Friday, 2020-07-03

*** elvispre <elvispre!> has joined #yocto00:06
*** yann <yann!> has quit IRC00:12
*** splatch <splatch!~splatch@> has quit IRC01:00
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC01:27
*** otavio <otavio!~otavio@> has joined #yocto01:29
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto01:29
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC01:43
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto01:45
*** dreyna <dreyna!> has quit IRC01:49
*** otavio <otavio!~otavio@debian/developer/otavio> has quit IRC01:50
*** otavio <otavio!~otavio@> has joined #yocto01:52
*** otavio <otavio!~otavio@debian/developer/otavio> has joined #yocto01:52
*** sgw_ <sgw_!> has quit IRC02:18
*** zkrx <zkrx!> has quit IRC02:37
*** zkrx <zkrx!> has joined #yocto02:44
*** rcw <rcw!> has quit IRC02:47
*** matthewzmd <matthewzmd!> has quit IRC03:48
*** matthewzmd <matthewzmd!> has joined #yocto03:48
*** micka <micka!> has quit IRC03:51
*** sakoman <sakoman!> has quit IRC04:27
*** micka <micka!> has joined #yocto04:32
*** paulg <paulg!> has quit IRC04:34
*** rcoote <rcoote!> has joined #yocto04:36
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has quit IRC04:43
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has joined #yocto04:44
*** ssajal <ssajal!> has quit IRC04:51
*** sgw2 <sgw2!> has joined #yocto05:01
*** sgw1 <sgw1!sgw@nat/intel/x-ozxhxlaxwlhgpjxy> has quit IRC05:02
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has quit IRC05:08
*** wallthar <wallthar!> has joined #yocto05:13
*** kroon <kroon!~kroon@> has joined #yocto05:16
*** agust <agust!> has joined #yocto05:19
*** gtristan <gtristan!~tristanva@> has quit IRC05:20
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has joined #yocto05:21
*** wallthar <wallthar!> has left #yocto05:21
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has quit IRC05:24
*** bluelightning <bluelightning!~paul@pdpc/supporter/professional/bluelightning> has quit IRC05:30
*** ssajal <ssajal!> has joined #yocto05:32
*** AndersD <AndersD!> has joined #yocto05:36
*** mranostay <mranostay!~mranostay@pdpc/supporter/active/mranostay> has joined #yocto05:48
*** pharaon2502 <pharaon2502!> has joined #yocto05:48
*** hard2b <hard2b!~hard2btru@> has quit IRC05:53
*** ssajal <ssajal!> has quit IRC05:56
*** sstiller <sstiller!> has joined #yocto05:59
*** pohly <pohly!> has joined #yocto06:01
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has joined #yocto06:02
*** jobroe <jobroe!> has joined #yocto06:08
*** leon-anavi <leon-anavi!~Leon@> has joined #yocto06:11
*** gtristan <gtristan!~tristanva@> has joined #yocto06:13
*** beratiks <beratiks!52de0992@> has joined #yocto06:25
*** gtristan <gtristan!~tristanva@> has quit IRC06:30
*** gtristan <gtristan!~tristanva@> has joined #yocto06:37
Letothe2ndgood morning new age dudes!06:40
PaowZmornin' Letothe2nd !06:40
PaowZtoday is friday..06:41
*** wallthar <wallthar!> has joined #yocto06:41
PaowZ..and I may start off with the first question of the day in this chan.. I'd need to display only one graphic application.. do I need a window manager like matchbox ? If no, how I can get rid of it so that I only keep xorg service onboard ? I played around with packagegroup-core-x11-xserver and packagegroup-core-x11-utils..with no luck so far..06:46
Letothe2ndPaowZ: it depends very much on the application, make can use fb or GL directly. thenyou don't need any wm06:49
PaowZhm.. I see.. moreoever, this is an openCV application.. with dependencies against gtk..06:51
Letothe2ndfor a first guess, i would take a look at whats inside those packagegroups, add the things you need directly to the RDEPENDS of your application and remove the packagegroup from the image.06:52
*** yacar_ <yacar_!> has joined #yocto06:53
PaowZI gave a try to remove matchbox-terminal and matchbox-wm and /usr/bin/Xorg completely vanished from rootfs..06:53
Letothe2ndrad again what i wrote.06:54
*** mckoan|away is now known as mckoan06:54
PaowZsorry, this is my poor emglish skills that make you think I didnt get your sentence.. plus, QA stuff bothered me around PN for RDEPENDS..06:56
*** yacar_ <yacar_!> has quit IRC06:58
PaowZwell.. this looks quite intricate..06:59
*** yacar_ <yacar_!> has joined #yocto06:59
Letothe2ndif something is unclear, i'm happy explain it. what i do not like is when i'm explaining something and then the one who asks just ignores what i said and does something entirely different and wonders why it still doesn't work. so, whats not clear about the suggestion?06:59
beratiksHi friends, I have a SOM based on imx8qxp with sumo branch yocto project. My boot process hanging (Because of systemd I think). For debug I opened systemd debug.And I saw that systemd stuck at loop. I copy my hanging systemd loop logs. Log : . Can you help me?06:59
*** lucaceresoli <lucaceresoli!> has joined #yocto07:02
*** polaris- <polaris-!> has joined #yocto07:03
Letothe2ndberatiks: without further digging into it: 1) start with what your som vendor provided. 2) if that doesn't even work, yell at them 3) once it works, start addind your modifications one by one until stuff breaks.07:04
PaowZLetothe2nd: Apologies if you felt ignored.. I wanted to append some more context.. I guess it's way beyond my whole comprehension of this gigantic stuff and my question is not just about a yes/no question..I'll sort this out. tks..07:05
PaowZI'll then rephrase for more specific question..07:06
Letothe2ndPaowZ: what wasn't targetted at you specifically, sorry if i sounded like it. i've just had a bunch of quite frustrating conversations lately. in a nutshell: pull the stuff you need from the packagegroup into your RDEPENDS, then remove the packagegroup. that is the classic approach07:06
*** hpsy <hpsy!~hpsy@> has joined #yocto07:07
*** seebs <seebs!~seebs@> has quit IRC07:07
beratiksLetothe2nd: problem that I use my manufacturer release project. After several continously reboot. My boot process hanging.07:08
Letothe2ndberatiks: go yell at the manufacturer. if they sell it, they shall support and debug it.07:09
beratiksLetothe2nd: ok thanks. I will force.07:11
*** DanielFuchs <DanielFuchs!> has joined #yocto07:11
*** DanielFuchs <DanielFuchs!> has left #yocto07:12
mihaiLetothe2nd, that's a long road :)07:13
*** DanielFuchs <DanielFuchs!> has joined #yocto07:15
PaowZthat sounds clear, Sepp. And I totally get the point of having to reply to some random dudes when it comes to answer vague/unclear questions.. however, I take it for myself as an improvement to make.. so I'll bear this in mind, I dont want to break the toy apart ;)07:15
*** DanielFuchs <DanielFuchs!> has left #yocto07:15
*** seebs <seebs!~seebs@> has joined #yocto07:16
Letothe2ndmihai: road to ruin?07:19
*** jobroe <jobroe!> has quit IRC07:19
*** xtron <xtron!~xtron@> has joined #yocto07:20
mihaiand despair07:20
Letothe2ndmihai: which of my statements are you referring to, exactly? ;-)07:20
*** jobroe <jobroe!> has joined #yocto07:20
mihaiyelling at the manufacturer07:21
Letothe2ndTBH, my take on the support given in this channel is that we help people in understanding stuff and getting their own tasks done. however if you buy something, and not even the default setup they hand out works, or is lacking documentation, then its their task to fix and support it. they got the money, after all.07:23
Letothe2ndand here are a couple of folks who will happily work for them and fulfill the task, if paid accordingly.07:23
Letothe2ndand any mfg who hands out sumo these days deserves to be yelled at in any case.07:24
*** rcoote <rcoote!> has quit IRC07:24
PaowZ..looks like you're in shape, this morning xD07:25
Letothe2ndnah, seriously. hardly anybody is doing stuff like this just for fun. most of us are doing this for a living in some form. so i see no reason to work for free or take away business from others, just that some cheapass manufacturer can have higher margins because cutting down on support.07:27
mihaiLetothe2nd, :)) I'm hoping that's just an older release that the custom is sticking to because of reasons07:27
*** nameclash <nameclash!> has joined #yocto07:29
Letothe2ndthen the mfg can accordingly respond: "here take this brand new shiny and supported BSP we have ready for you right here", and all is well :)07:29
Letothe2ndit might be just me, but i have strong opinions on sorting out other peoples crap (e.g. working for free)07:30
*** xtron1 <xtron1!~xtron@> has joined #yocto07:31
mihaifor some mfg a yocto BSP is at a "proof of concept" level, something to get you started and for them to tick in a checklist when releasing new products07:32
Letothe2ndyup. but thats their problem, not mine.07:33
Letothe2ndand unless we handle their customers for them, they'll just go on like that.07:33
mihaiafter that, although they aspire to having community presence and support, their involvement is somewhere at the bottom of their list of priorities07:33
*** xtron <xtron!~xtron@> has quit IRC07:33
Letothe2nds/unless/as long as/g07:33
Letothe2ndexactly, and thats why i do not support such cases.07:34
*** nameclash is now known as mbulut07:34
*** fbre <fbre!91fdde45@> has joined #yocto07:35
mihaiand you're not actually doing it for free, I get the feeling you're more of enjoying this :)07:35
Letothe2ndit depends. it has various political reasons that this is the form of giving-back that i do. and i am a very social person, thats true.07:36
mihaiit's a "spiritual" gain there somewhere :))07:36
fbreHi, as I tried to mount overlayfs I got this error: overlayfs: upper fs does not support xattr.    Any idea what is missing in my yocto distro?07:36
*** mbulut is now known as mbulut_nameclash07:36
mihaifbre, xattr :)07:36
*** mbulut_nameclash is now known as mbulut07:36
mihaiin DISTRO_FEATURES07:37
fbrereally that easy solution or is it a joke? :)07:37
Letothe2ndoO( when did we ever joke in here? )07:38
mihainot a joke, smiling because the easy solution as you said07:39
fbrehehe OK '=D  thanx. Sometimes it's so easy that one cannot see the forest because of the trees (as we say in German)07:39
*** mbulut is now known as mbulut_nameclash07:39
* Letothe2nd notes that the docs might need updating?
polaris-hi, I want to apply a config fragment to the kernel config. Seems bitbake already "sees" the fragment file because it rebuilds the kernel when I edit options in the file. However, the option does not seem to be "merged" to the final .config file and the feature I want to have (overlayfs) is not available on the final image.07:40
mihaiLetothe2nd, yup, there are quite a bunch of them not in there07:40
mihaifbre, nice saying07:41
qschulzpolaris-: inheriting kernel-yocto in your kernel recipe?07:43
qschulzpolaris-: that was a question actually.. are you inheriting kernel-yocto in the recipe?07:43
polaris-qschulz: yes, I changed that yesterday and it "improved" things a bit :)07:43
polaris-qschulz: before that change bitbake ignored the fragment07:44
qschulzpolaris-: how did you create your config fragment?07:44
polaris-qschulz: as described here:
polaris-qschulz: eventually running "bitbake linux-yocto -c diffconfig"07:45
polaris-well, not linux-yocto in my case07:45
qschulzpolaris-: I like people following the docs :) thank you.07:45
polaris-qschulz: I try my best :)07:46
qschulzwhere is your config fragment?07:46
qschulzin your layer?07:46
polaris-qschulz: yes, I have a layer .. currently only with the bbappend and the fragment.cfg.07:46
*** yann <yann!> has joined #yocto07:47
polaris-qschulz: created using bitbake-layers07:47
fbrethis way?:    DISTRO_FEATURES_append ?= "xattr"07:47
fbreI wonder what xattr actually does07:48
polaris-qschulz: the .bbappend only contains the two lines FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:" and SRC_URI += "file://overlayfs.cfg"07:49
qschulzfbre: _append requires in 99% of cases a leading space ;) No need for ?=, a simple = is fine07:51
fbreqschulz: thanx (y)07:52
mihaifbre, it enables xattr support in some of the packages, the ones that packageconfig it anyway, you can also check if there are xattr kernel config you're missing in case of "exotic layering", on top of e.g. tmpfs07:56
fbreyes, I first tried a test with overlayfs by mounting directories I created in /tmp07:59
*** Ox861726f6c64 <Ox861726f6c64!~Ox861726f@gateway/tor-sasl/ox861726f6c64> has joined #yocto08:00
mihaifbre, you can check the running config /proc/config.gz if you still have it on08:04
*** Dracos-Carazza <Dracos-Carazza!> has quit IRC08:19
*** Dracos-Carazza <Dracos-Carazza!> has joined #yocto08:19
*** ruz <ruz!> has joined #yocto08:20
ruzMy bitbake parser doesn't translate calls like ${@bb.utils.contains} in a specific shell task.08:27
ruzSo when this task is executed, I have a 'bad substitution' error08:27
Letothe2ndruz: example on a pastebin, please?08:28
*** BloodSurfer <BloodSurfer!~BloodSurf@2a02:8070:181:d000:f145:e69a:21b5:90d5> has joined #yocto08:32
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has joined #yocto08:32
*** Bunio_FH <Bunio_FH!> has joined #yocto08:36
BloodSurferHeyho guys. I'm trying to build and add a signed fitimage for a imx6 board based on phytec's SoM to the /boot directory. I can't quite wrap my head around the installation process of a bbclass. The do_deploy task does create symlinks etc in the ${TOPDIR}/deploy folder. How can I install the symlink to the final destination. The do_install() and FILES_${PN} is ignored afaik08:36
BloodSurferDoes anyone have any suggestions how to proceed?08:36
*** ecksun <ecksun!> has joined #yocto08:39
Letothe2ndBloodSurfer: not sure if you're not thinking too complicated? if you install the kernel package in your image, it should end up exactly there. so its more the task of having the recipe generate the correct artifact.08:40
*** Emantor_ <Emantor_!> has quit IRC08:42
*** mbulut_nameclash <mbulut_nameclash!> has quit IRC08:42
*** mbulut_nameclash <mbulut_nameclash!> has joined #yocto08:42
*** Emantor <Emantor!> has joined #yocto08:43
ruzLetothe2nd here:
Letothe2ndruz: hum, are you sure its not a classic version mismatch problem? you state building for thud, but the snippet you link seems to be rather new. so maybe the bb.utils semantics have changed (just guessing)08:49
BloodSurfer@Letothe2nd: As far as I understand this class it does not build but only package the zImage and I need to modify it to add the keys for the signature08:49
BloodSurferbtw. here's the link to the class:
BloodSurferI assume this line kills the installation process: 31 do_install[noexec] = "1"08:51
Letothe2ndBloodSurfer: well the comments at the top pretty much state what the class expects. but as this is their custom, special sauce, i would actually point to phytec for also supporting it.08:51
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has joined #yocto08:53
BloodSurfer@Letothe2nd: Yep, the comment is actually why I ended up writing my own fitimage recipe inheriting their stuff08:54
ruzLetothe2nd: I don't think, because when I execute 'bitbake machine-image' ("default" image), I don't have the error, the function/task do_makesystem is correctly parsed. My image is very trivial: it requires
ruzNotice that others functions of file '' are parsed. Only 'do_makesystem' is not parsed08:55
Letothe2ndruz: ok, no idea then. sorry.08:55
Letothe2ndBloodSurfer: and their own class works if used (crrectly)?08:57
BloodSurfer@Letothe2nd: Couldn't verify yet since I'm trying to this exactly ;)08:58
qschulzruz: know that qualcomm have **many** hacks and change **a lot** of things from Yocto "internals". So you're in for a treat :) Wild guess, but is it possible IMAGE_FSTYPES does not exist in your image?09:02
qschulzruz: bitbake <image> -e | grep -e "^IMAGE_FSTYPES" does this return something?09:02
ruzYes, and "funfact" as soon as I use another image than theirs one, I have the issue09:06
ruzWhere can I find informations of bitbake parser? Is there alternative of 'bitbake -DDD' to debug parser?09:06
*** DanielFuchs <DanielFuchs!> has joined #yocto09:11
*** wertigon <wertigon!> has joined #yocto09:14
qschulzruz: check what they're doing in the final image that differs from the base-image. Some additional classes, some additional/changed variables/tasks etc...09:18
qschulzruz: the code for the bb.utils.contains is in poky/bitbake/lib/bb/utils.py09:19
*** ruz <ruz!> has quit IRC09:20
wertigonHi, getting a compile/configuration error with gstreamer in poky, see
wertigonComplaining about gstreamer-gl not found09:21
wertigondunfell branch09:21
wertigonCould be meta-openembedded instead of poky though09:21
*** ruz <ruz!> has joined #yocto09:22
ruz@qschulz you're right, I found differences of existing variables between 'machine-image' and my own image (based on first one). I found a .bbappend of 'machine-image', but it seems it is not appended with my image. If my image does a require of 'machine-image', .bbappend of 'machine-image' should works, right?09:34
*** beratiks <beratiks!52de0992@> has quit IRC09:36
qschulzruz: no09:38
mihairuz, somehow the parser is not catching that as python code and reaches the shell09:50
mihairuz, how are you defining that function, can you pastebin that?09:51
mihairuz, it's the same link :)09:57
*** ruz <ruz!> has quit IRC10:04
*** BobPungartnik <BobPungartnik!~BobPungar@> has joined #yocto10:09
fbremihai: config /proc/config.gz   leads to -sh: config: not found10:15
Letothe2ndfbre: zcat or zgrep it10:15
fbrethe xattr error is still around with the new yocto10:15
fbreI mean my new yocto build10:15
fbrezcat /proc/config.gz | grep xattr    is empty10:17
*** JEEB <JEEB!> has joined #yocto10:17
fbreDid DISTRO_FEATURES_append = " xattr"  not work?10:18
Letothe2ndfbre: where did you add that line?10:18
*** JEEB <JEEB!> has quit IRC10:18
*** JEEB <JEEB!~jeeb@unaffiliated/jeeb> has joined #yocto10:18
*** JEEB <JEEB!~jeeb@unaffiliated/jeeb> has left #yocto10:19
*** fbre <fbre!91fdde45@> has quit IRC10:20
*** nagua <nagua!> has joined #yocto10:34
*** wallthar <wallthar!> has quit IRC10:46
*** ruz <ruz!> has joined #yocto10:48
ruz@mihai Sorry, what function are you meaning? The link contains do_makesystem definition (with original file link), how it is interpreted (run.do_makesystem) and where it is added as task (qimage class)10:48
kanavin_homeRP: rburton: I will be in the lake cabin for the next month, email will work as usual, but irc most likely not10:52
Letothe2ndkanavin_home: enjoy!10:53
RPkanavin_home: thanks for the heads up, enjoy it! I could do with escaping to a lake cabin atm :)10:53
RPkanavin_home: should we plan for an AUH run at some point?10:54
RPkanavin_home: been appreciating the patch review/email help btw, thanks!10:54
kanavin_homeRP: AUH will run as scheduled on the 15th, no?10:58
kanavin_homeor is it still not on a schedule?10:58
kanavin_homeI do want to see if sending AUH emails to oe-core will broaden the patch submissions10:59
*** sh00p <sh00p!> has quit IRC11:00
*** jofr <jofr!> has quit IRC11:00
*** polaris- <polaris-!> has quit IRC11:01
*** lucaceresoli <lucaceresoli!> has quit IRC11:09
*** fbre <fbre!91fdde45@> has joined #yocto11:11
*** behanw <behanw!uid110099@gateway/web/> has quit IRC11:11
fbremihai: Could you please tell that command again for the /proc directory? I lost that irc chat line due page reload of the browser11:14
fbresomehow with cat or something11:14
mihaifbre, I was telling you about checking the running kernel configuration in /proc/config.gz11:15
mihaifor xattr on tmpfs11:15
mihairuz, right, sorry, I haven't scrolled all the way down11:15
Letothe2ndfbre: you can use zcat or zgrep to work on that config.gz - and, where did you set DISTRO_FEATURES_append?11:16
*** rob_w <rob_w!~bob@unaffiliated/rob-w/x-1112029> has quit IRC11:17
yannwith hashequiv disabled, I still have a few spurious rebuilds - the good news is they are much easier to track now :)11:18
fbrezcat /proc/config.gz | grep xattr    is still empty :(11:19
mihaifbre: grep -i, or XATTR11:19
* Letothe2nd really doesn't want to ask a thrid time.11:20
kpo_hello, I'm writing my own do_fetch() function i which I download a pack using wget (I'm using custom do_fetch, because I have to send some --post-data to accept license agreement). As long as I am in do_fetch() function, I can see downloaded tar in ${S}, but the same path in do_unpack or do_install is empty. Should I mark the file somehow that it is not deleted between those functions?11:20
fbreI have Linux\trunk\Yocto\Sumo\sources\meta-fbre\conf\distro\fsl-imx-wayland-fbre.conf     where I added xattr to that already existing line: DISTRO_FEATURES_append = " wayland pam systemd opengl vulkan xattr"11:20
fbreShould work, shouldn't it?11:21
Letothe2ndfbre: yes, i just wanted to check if you're not setting it in the image recipe11:21
fbreroot@imx8fbre:/tmp# zcat /proc/config.gz | grep XATTR# CONFIG_EXT2_FS_XATTR is not set# CONFIG_TMPFS_XATTR is not set# CONFIG_JFFS2_FS_XATTR is not set# CONFIG_SQUASHFS_XATTR is not set11:22
yannone looks strange at first sight: linux-firmware do_package hash changes because of RK_LICENSE_PATH = "${LAYERDIR}/files/custom-licenses" in the meta-rockchip vendor layer.  This ends up in LICENSE_PATH.  For some reason _that one_ (and not all other layers using LICENSE_PATH += "${LAYERDIR}/licenses") impacts the hash.  Isn't that strange ?11:23
fbresorry, the newlines got lost during copy'n'paste11:23
yannfbre: that's what pastebin is for :)11:24
mihaiI wonder if there's a kernel config fragment that enable those11:25
yannah, LICENSE_PATH is in BB_HASHBASE_WHITELIST and that does not extend to variables used to build it.  Shouldn't this at least give a warning ?11:28
*** mbulut_nameclash <mbulut_nameclash!> has quit IRC11:33
*** wooosaiiii <wooosaiiii!> has joined #yocto11:37
*** wooosaiiii <wooosaiiii!> has quit IRC11:39
*** BobPungartnik <BobPungartnik!~BobPungar@> has quit IRC11:43
*** wooosaiiii <wooosaiiii!> has joined #yocto11:44
*** wooosaiiii <wooosaiiii!> has joined #yocto11:45
*** wooosaiii <wooosaiii!> has joined #yocto11:45
*** wooosaiiii <wooosaiiii!> has quit IRC11:47
*** berton <berton!~berton@> has joined #yocto11:48
*** gtristan <gtristan!~tristanva@> has quit IRC12:03
*** mbulut_nameclash <mbulut_nameclash!> has joined #yocto12:07
*** polaris- <polaris-!> has joined #yocto12:07
kpo_FYI, got what I had wrong - I should've used DL_DIR, it keeps contents, and second (or first of all), I could've used FETCHCMD_wget and added my --post-data there ;)12:09
*** gtristan <gtristan!~tristanva@> has joined #yocto12:22
fbrehmmm, there are a few kernel config settings for XATTR12:35
fbrefor each filesystem12:36
fbre... going to build a new kernel. Happy weekend all! See you12:36
*** pohly <pohly!> has quit IRC12:37
*** fbre <fbre!91fdde45@> has quit IRC12:38
*** Ox861726f6c64 <Ox861726f6c64!~Ox861726f@gateway/tor-sasl/ox861726f6c64> has quit IRC12:40
*** xtron1 <xtron1!~xtron@> has quit IRC12:45
*** paulg <paulg!> has joined #yocto12:52
*** dmoseley <dmoseley!~dmoseley@> has quit IRC12:59
*** dmoseley <dmoseley!~dmoseley@> has joined #yocto13:02
*** learningc <learningc!~pi@> has quit IRC13:06
*** DanielFuchs <DanielFuchs!> has left #yocto13:07
*** dmoseley <dmoseley!~dmoseley@> has quit IRC13:07
*** polaris- <polaris-!> has quit IRC13:10
*** maudat <maudat!> has joined #yocto13:12
*** ssajal <ssajal!> has joined #yocto13:13
JPEWblerg, I made a mess of archiver.bbclass. Server me right for trying to fix a race condition :)13:16
JPEWTook 3 patches to add "after do_preconfigure" to a task13:17
JPEWRP: Do we test archiver.bbclass on the AB? I'm a little suprised it didn't catch anything13:19
RPJPEW: yes, we do13:19
RPJPEW: meta/lib/oeqa/selftest/cases/archiver.py13:19
RPclearly not well enough13:20
*** dmoseley <dmoseley!~dmoseley@> has joined #yocto13:20
*** xtron1 <xtron1!~xtron@> has joined #yocto13:20
JPEWRP: My first guess on a quick review is that it's only testing very specific recipes, and the gcc source is not one of them13:21
RPJPEW: sounds about right13:22
JPEWI think adding some tests would be easy.... they should go fast b/c of sstate13:24
RPJPEW: that sounds good :)13:24
* RP is getting tired of chasing races :/13:24
*** dmoseley <dmoseley!~dmoseley@> has quit IRC13:26
*** nagua <nagua!> has quit IRC13:28
*** Bunio_FH <Bunio_FH!> has quit IRC13:28
*** dmoseley <dmoseley!~dmoseley@> has joined #yocto13:29
*** kroon <kroon!~kroon@> has quit IRC13:31
*** ruz <ruz!> has left #yocto13:31
*** dmoseley_ <dmoseley_!~dmoseley@> has joined #yocto13:37
*** dmoseley <dmoseley!~dmoseley@> has quit IRC13:37
*** grumble <grumble!~Thunderbi@freenode/staff/grumble> has quit IRC13:43
*** grumble <grumble!~Thunderbi@freenode/staff/grumble> has joined #yocto13:43
*** matthewzmd <matthewzmd!> has quit IRC13:47
*** matthewzmd <matthewzmd!> has joined #yocto13:48
yannhm, touching BB_HASHBASE_WHITELIST (to workaround that RK_LICENSE_PATH causes "metadata is not deterministic", what's the problem ?  Shouldn't BB_HASHBASE_WHITELIST list itself ?13:50
*** mbulut_nameclash is now known as mbulut13:52
*** lucaceresoli <lucaceresoli!> has joined #yocto13:53
*** sajjad__ <sajjad__!~xtron@> has joined #yocto14:01
*** AndersD <AndersD!> has quit IRC14:01
*** sstiller <sstiller!> has quit IRC14:01
*** xtron1 <xtron1!~xtron@> has quit IRC14:03
*** sakoman <sakoman!> has joined #yocto14:04
*** matthewzmd <matthewzmd!> has quit IRC14:04
*** matthewzmd <matthewzmd!> has joined #yocto14:04
RPyann: nothing that references BB_HASHBASE_WHITELIST is in the code that gets checksummed14:06
yannThen why would BB_HASHBASE_WHITELIST += "RK_LICENSE_PATH" wreak havoc like that ?  Virtually all tasks suddenly trigger a  "basehash changed"14:10
yann(I had added it to my own layer instead of modifying meta-rockchip, but I don't think it would have any impact)14:11
*** jobroe <jobroe!> has quit IRC14:13
RPyann: was that in layer.conf?14:14
RPyann: I had this discussion with jonmason and rburton yesterday :/14:14
RPyann: if it was layer.conf, the issue is the += masks out the ?= from bitbake.conf14:15
yannyes, that was in my own layer's layer.conf14:15
yannyou mean, "+=" does not play well with "?=" ?14:17
qschulzyann: += override ?= if += is "resolved"/parsed before ?=14:18
RPyann: its all about ordering. layer.conf is parsed really early. If you set there, it has a value by the time bitbake.conf is parsed. ?= means set if not already set, its already set so it doesn't get changed14:18
yannoh fck :|14:18
qschulzyann: you could use _append instead in your layer.conf (don't know if it's proper but it'll probably work)14:18
RPright, use _append14:18
yannwell, I suggested them to just avoid using an intermediate variable referencing ${LAYERDIR}14:20
RPyann: that seems sensible14:24
matthewzmdwhy does some recipes are LICENSED under GPL but has a commercial LICENSE_FLAGS?14:29
jonmasonThe "solution" we had was that it needs a _append not "+=" due to the order of parsing14:30
jonmasonI fixed up that patch and shamed the person who wrote it  :)14:31
RPmatthewzmd: commercial could be patents14:32
RPjonmason: I've mailed the arch list as I think there is a bigger problem14:32
*** maudat <maudat!> has quit IRC14:33
*** maudat <maudat!~moda@> has joined #yocto14:33
jonmasonRP: yes, it seems like something that needs to be handled better than "use _append"  :)14:34
RPkergoth: any thoughts on that problem? (default values for vars)14:34
RPjonmason: I'm worried when people like ross don't understand the assignments :/14:34
RPjonmason: ross did briefly try and convince me check-layer was broken! :)14:35
jonmasonRP: devday, plus 11pm your time, plus having to explain it to me  :)14:35
RP(which sadly I could believe)14:35
jonmasonRP: it is broken.  I've been complaining to him all week14:35
jonmasonall he tells me is to fix it ;-)14:35
RPwell, it was right about this ;-)14:35
RPjonmason: send patches!14:36
RPjonmason: sadly I have a long list of other things ross tells me are broken to work through ;-)14:36
jonmasoncheck-layer fails on meta-oe.  So no one must be including it when they run it, which has me scared.14:36
RPjonmason: I've wanted to add that to the YP AB checks14:37
jonmasonthe last bit I'm hacking through is meta-virt (due to meta-arm-autonomy), and it is a horror story14:37
RPjonmason: lack of people to help fix any issues is the blocker14:37
RPcan't get clean builds now14:37
jonmasonI'm adding check-layer to the regressions I run on meta-arm patches.  So once it is working, it'll never break again (in theory)14:38
*** mbulut <mbulut!> has quit IRC14:39
*** matthewz_ <matthewz_!> has joined #yocto14:39
*** matthewzmd <matthewzmd!> has quit IRC14:40
jonmasonI'll try to get my "fixes" out to the list14:41
*** matthewz_ is now known as matthewzmd14:45
matthewzmdRP: the recipe-in-queston is - i want to avoid LICENSE_FLAGS_WHITELIST if possible, and i don't see why it has to be commercial14:48
*** dmoseley <dmoseley!~dmoseley@> has joined #yocto14:49
smurraymatthewzmd: AAC is patent encumbered, AFAIK14:50
*** dmoseley_ <dmoseley_!~dmoseley@> has quit IRC14:50
smurraymatthewzmd: see
*** florian <florian!~florian_k@Maemo/community/contributor/florian> has quit IRC14:55
matthewzmdsmurray: okay :-(14:58
*** rcoote <rcoote!~rcoote@2a02:908:694:300:6835:4214:6819:3ca> has joined #yocto14:59
armpitjonmason, I have sent fixes to meta-oe to help it pass but the community keeps breaking it with new packages etc.15:07
armpitits a loosing cycle15:08
jonmasonarmpit: getting check-layer as part of some kind of patch acceptance test would be a good start.  It takes almost no time to run15:08
jonmasonas soon as I get it working on meta-arm, I'll be running it for every patch there15:09
armpitI don't disagree.15:09
armpitI wonder if we code get the check added to the meta-oe build in the AB15:11
armpits/code/could/ ..... still waking up15:12
jonmasonarmpit: I think that is what RP was saying above15:12
armpitIll send a patch then15:12
jonmasoncool, then you can fix all the breaks I'm seeing :)15:13
armpitsure.. I can certainly help15:13
* armpit better check dunfell15:13
jonmasonLAYERDEPENDS_meta-python is hosed because "(>= 12) " doesn't parse15:13
armpitthere is a fix for that in master and I think is maybe in the queue for dunfell core15:14
jonmasonand I think needs "RPROVIDES =  'apache'"15:14
armpitnico sent the fix15:14
jonmasonI could be stale a day or two15:14
armpitjonmason, its in the dunfell-next so in the pipeline for merging15:15
jonmasonI'm just running master right now15:16
jonmasontoo many combinations15:16
armpithehe, been there15:16
*** ruz <ruz!> has joined #yocto15:16
jonmasonI thought it would take 1-2 days to get it working, boy was I wrong15:16
armpitthen spread that across different build systems15:16
zeddiiheh. I hear you. I've been fixing meta-virt ones, and it takes a while for a single run on my builder.15:17
zeddiiwhich day do you guys get off for the 4th ? Monday ?  since you are working today ?15:17
paulgthinking it is the monday.15:18
armpitfor me it was yesterday and today15:18
zeddiiyou guys spend your day off ... 'oddly'15:18
armpitvolunteering does not have holidays15:19
jonmasonzeddii: I remember you working on the 1st15:19
zeddiiyou may have thought that, but I spent about 45 minutes total in sessions that day :d15:19
zeddiiI blame the canadian hating LF schedulers for that.15:19
armpitnow the truth comes out15:19
jonmasonwhy did it move a week later, so dumb15:19
armpita holiday week to boot15:20
*** NiksDev2 <NiksDev2!~NiksDev@> has joined #yocto15:22
*** leon-anavi <leon-anavi!~Leon@> has quit IRC15:23
*** kanavin_home <kanavin_home!~ak@2a02:2450:1011:56f:b652:c378:b428:fdf> has quit IRC15:24
*** halstead_ <halstead_!> has joined #yocto15:25
*** blueness_ <blueness_!~blueness@gentoo/developer/blueness> has joined #yocto15:25
*** warthog19 <warthog19!> has joined #yocto15:25
*** adelcast <adelcast!> has quit IRC15:25
*** blueness <blueness!~blueness@gentoo/developer/blueness> has quit IRC15:25
*** NiksDev <NiksDev!~NiksDev@> has quit IRC15:25
*** nerdboy <nerdboy!~sarnold@gentoo/developer/nerdboy> has quit IRC15:25
*** halstead <halstead!> has quit IRC15:25
*** u1106 <u1106!> has quit IRC15:25
*** warthog9 <warthog9!> has quit IRC15:25
*** adelcast <adelcast!~adelcast@2605:6000:101c:3b5:c255:6326:5d60:1a29> has joined #yocto15:25
*** warthog19 is now known as warthog915:26
*** u1106 <u1106!> has joined #yocto15:26
*** lucaceresoli <lucaceresoli!> has quit IRC15:28
*** thaytan <thaytan!> has quit IRC15:29
*** thaytan <thaytan!> has joined #yocto15:30
ruzqschulz mihai finally I found the solution! Qualcomm has many 'machine-image.bbappend', and because I required from '', all of these *.bbappend weren't added. So I have added (at hand, ugly...) the Qualcomm BSP .bbappend as include file for my image, and it works! But I don't know why it was parsing error for do_makesystem task...15:31
qschulzruz: I heavily do not recommend doing that... Bbappends are "applied" in a specific order that depends on layer priorities and the order in bblayers.conf... It's highly possible you got the order wrong and might wonder later why it does not work (anymore)15:33
qschulzI'd try to find what was needed (add bbappends one by one and find the one that fixes it)15:33
*** hpsy <hpsy!~hpsy@> has quit IRC15:36
ruzqschulz I added only at hand the right bbappend. For now, I didn't add others ones.15:40
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has joined #yocto15:41
qschulzruz: better than what I imagined :)15:43
*** Ox861726f6c64 <Ox861726f6c64!~Ox861726f@gateway/tor-sasl/ox861726f6c64> has joined #yocto15:43
*** lucaceresoli <lucaceresoli!> has joined #yocto15:49
champagneghi! is there a way for a package to be a subset of another one or to partially install a package? My specific use case is with mesa-megadrivers, i'm trying to install only the .so for my specific drm driver.15:51
qschulzchampagneg: no, a file can only be part of one package15:51
qschulzchampagneg: however, you can say that one package depends at runtime (or recommends) on another one with RDEPENDS_<package_name> (respectively RRECOMMENDS_<package_name>)15:52
qschulzchampagneg: so if you want just a subset, a technique could be to make that file(s) part of a new package that is then added to RDEPENDS of the package that originally had this file15:53
*** pharaon2502 <pharaon2502!> has quit IRC15:55
champagnegqschulz: ok, thx. If the parent packages uses globbing, my file would get picked up anyway, right? There's no way to use FILES_xyz_remove after globbing occured?15:55
*** sajjad__ <sajjad__!~xtron@> has quit IRC15:55
qschulzchampagneg: add your new package BEFORE the original package15:58
qschulzchampagneg: the leftmost package which has a FILES_${PN} matching for your file will get the file and not the others15:59
*** sajjad__ <sajjad__!~xtron@> has joined #yocto15:59
champagnegqschulz: ah OK, i see. Thanks for the tips!15:59
qschulzchampagneg: my pleasure :)16:00
*** goliath <goliath!> has joined #yocto16:07
*** polaris- <polaris-!> has joined #yocto16:08
*** T_UNIX <T_UNIX!uid218288@gateway/web/> has quit IRC16:12
*** polaris- <polaris-!> has quit IRC16:13
*** mckoan is now known as mckoan|away16:23
*** ruz <ruz!> has left #yocto16:23
*** sgw1 <sgw1!~sgw@> has joined #yocto16:29
*** sajjad__ <sajjad__!~xtron@> has quit IRC16:40
*** lucaceresoli <lucaceresoli!> has quit IRC16:58
*** RobertBerger <RobertBerger!> has joined #yocto17:09
*** goliath <goliath!> has quit IRC17:09
*** nerdboy <nerdboy!~sarnold@> has joined #yocto17:10
*** nerdboy <nerdboy!~sarnold@gentoo/developer/nerdboy> has joined #yocto17:11
*** pohly <pohly!> has joined #yocto17:24
RPzeddii: what is this "day off" thing?17:29
*** maudat <maudat!~moda@> has quit IRC17:32
*** maudat <maudat!> has joined #yocto17:34
zeddiiyou have to get a permission slip from bigfoot.17:34
paulgI hear he moved to Loch Ness to get away from US CV-19 outbreak.17:47
*** rcoote <rcoote!~rcoote@2a02:908:694:300:6835:4214:6819:3ca> has quit IRC18:24
RPpaulg: that isn't too far from here :)18:29
*** pharaon2502 <pharaon2502!> has joined #yocto18:32
*** rcoote <rcoote!~rcoote@> has joined #yocto18:40
*** pharaon2502 <pharaon2502!> has quit IRC18:45
paulgbetter go over and check that he did a 14d self-isolation.18:46
*** wertigon <wertigon!> has quit IRC18:47
*** halstead_ is now known as halstead18:58
*** rcoote <rcoote!~rcoote@> has quit IRC19:00
*** rcoote <rcoote!~rcoote@2a02:908:694:300:6835:4214:6819:3ca> has joined #yocto19:00
*** goliath <goliath!> has joined #yocto19:03
*** polaris- <polaris-!> has joined #yocto19:08
*** polaris- <polaris-!> has quit IRC19:13
*** NiksDev2 <NiksDev2!~NiksDev@> has quit IRC19:16
*** NiksDev2 <NiksDev2!~NiksDev@> has joined #yocto19:17
*** khem <khem!~khem@unaffiliated/khem> has quit IRC19:46
*** sajjad__ <sajjad__!~xtron@> has joined #yocto19:50
*** khem <khem!~khem@unaffiliated/khem> has joined #yocto19:50
*** sajjad__ <sajjad__!~xtron@> has quit IRC19:58
*** BCMM <BCMM!~BCMM@unaffiliated/bcmm> has quit IRC20:03
*** rcoote <rcoote!~rcoote@2a02:908:694:300:6835:4214:6819:3ca> has quit IRC20:03
*** rcoote <rcoote!~rcoote@> has joined #yocto20:04
*** Ox861726f6c64 <Ox861726f6c64!~Ox861726f@gateway/tor-sasl/ox861726f6c64> has quit IRC20:16
*** dakhouya <dakhouya!> has joined #yocto20:25
*** rcoote <rcoote!~rcoote@> has quit IRC20:32
*** splatch <splatch!~splatch@> has joined #yocto20:34
splatchhey folks, I'm back just to say thank you for assistance you gave me yesterday, one brooken night, couple of swearword20:36
splatchbut it started to work20:36
*** pohly <pohly!> has quit IRC20:43
smurraysplatch: good to hear!20:51
paulgif you aren't swearing, you aren't trying hard enough.   ;-)20:52
*** maudat <maudat!> has quit IRC21:00
splatchthat's correct, with tips from you smurray and JPEW I managed to get qemu working. All in all raw qemu didn't work for me, but I managed to get iso image mounted into virtualbox and tested there.21:03
splatchnow I've ran into a conflicts - wireguard-client-config-1.0.1-r1.core2_64 vs wireguard-tools-0.0.20190123-r0.core2_64 which both try to owe /etc/wireguard.21:05
splatchI know that pacman under arch has conflicts section for recipes which do step on each other - but in this particular case these seems to be close to each other..21:06
splatchis there a way to find sources of other layer on the disk?21:09
*** yacar_ <yacar_!> has quit IRC21:15
smurraysplatch: what package a file is coming from, you mean?21:16
smurraysplatch: oe-pkgdata-util <filename>, where filename is the path in the root fs, will tell you what package it comes from21:17
splatchsmurray: I am looking for conents of /etc/wireguard to see if I actually override it or not21:19
*** berton <berton!~berton@> has quit IRC21:19
smurraysplatch: you should be able to to that with oe-pkgdata-util, I think21:19
smurraysplatch: I don't see wireguard-client-config in the layer index, so my guess is it'd be doing something odd21:20
splatchwhen I skip my client config it goes just fine, but I dont see source file21:20
splatchsmurray: yup, client config is an addon of mine to get wireguard interface setup after boot21:21
smurraysplatch: if wireguard-client-config needs to replace a file that is usually in wireguard-tools, you'll need a wireguard-tools bbappend to rm the file in do_install_append21:22
smurraysplatch: two packages cannot provide the same file21:22
*** aidanh_ <aidanh_!~aidanh@unaffiliated/aidanh> has joined #yocto21:23
smurraysplatch: the cleaner way would be to split out a config package in the upstream wireguard-tools recipe with the affected file(s), so that that config package can be replaced more easily21:23
splatchI guess it would, I source recipe from very few samples available at github:
*** aidanh <aidanh!~aidanh@unaffiliated/aidanh> has quit IRC21:26
*** aidanh_ is now known as aidanh21:26
splatchok, seems the problem is actually this line
splatchthe directory is created already by tools so client should (can?) rely on its existence21:28
smurraysplatch: that's a bit surprising, I didn't think directories could conflict, only files.  Unless that behavior differs across the package manager options21:32
smurraysplatch: does the error message specifically mention /etc/wireguard?21:32
splatchit does: Error: Transaction check error: file /etc/wireguard conflicts between attempted installs of wireguard-client-config-1.0.1-r1.core2_64 and wireguard-tools-0.0.20190123-r0.core2_6421:33
splatchI've checked build contents and both wireguard-tools and client-config have /etc/wireguard as a directory21:34
splatchpardon, find . -name '*wireguard*' -type f|grep -i etc|sort shows only my recipe21:36
splatchhm, still all of them seem be valid dir and file entrie21:38
splatchmaybe that's the cause:
smurraysplatch: that only has an effect if there's actually things being installed by the wireguard-tools into that directory21:43
splatchsmurray: tools seems to ship empty /etc at least from what I see in the package output, anyhow will try with append you suggested21:45
smurraysplatch: from looking at src/Makefile, the issue might be that wireguard-tools explicitly creates the dir as 0700, and that wireguard-client-config recipe does not.  The package manager might see that as a conflict21:48
splatchshall I try with same chmod?21:48
smurraymight be interesting to see if that fixes it21:49
splatchmy pleasure :)21:49
*** pharaon2502 <pharaon2502!> has joined #yocto21:49
smurrayotherwise, rm it in a bbappend, I guess, though you'll want to dig around to see if that dir does need to be 0700.  I suspect it's recommended since I imagine there's keys in the config files...21:50
splatchthat's correct, it holds some secrets21:51
splatchI'd consider moving the key anyway to data partition later on, but so far I am in the stage of "lets see if that github thing works"21:52
splatchsmurray: you was right, its was a chmod issue, problem solved with change to client recipe. install -d -m 0700 ${D}${sysconfdir}/wireguard let build pass21:54
smurrayI could see the init script addition being useful in the wireguard-tools recipe, it's a bit surprising there's not an init script in the contrib directory21:54
smurraysplatch: cool.  Definitely annoying that the error message didn't highlight that21:54
splatchsmurray: that's third hardest thing in the programming nobody says outloud ;-)21:55
smurrayheh, yes21:55
smurrayI suspect in this instance, it's coming from the package manager and not OE, which is harder to get fixed if it's e.g. rpm21:56
splatchsmurray: I guess that wireguard has still smaller community than ovpn due to its popularity, hence there is fewer complainers (such me) and even less contributors capable of making patch.21:57
splatchsmurray: if package manager splits 900+ packages when it fails then you're right, it was him21:58
splatchwhat initially mislead me was beginning of the error "Could not invoke dnf."21:59
splatchanyhow, note to self - verify permissions when package manager complains about overriden file21:59
splatch[or dir]21:59
smurrayyeah, sometimes packaging errors in rootfs creation take some decoding.  If you're new to OE, it may not be immediately obvious that the process is to build packages, then use a package manager to build the image from them22:01
splatchwhat bothers me more are inotify constraints ;-) I can't search from any desktop text editor. Any attempt to open anything results in 300k entries which hang half of the system22:05
splatchso I can't quickly search for text occurances or patterns22:05
splatchbut that's might be me being used to IDE style work22:05
*** splatch <splatch!~splatch@> has quit IRC22:06
*** splatch <splatch!~splatch@> has joined #yocto22:06
* splatch wrote exit in wrong window22:07
smurrayI avoid searching down in the build output if at all possible, usually tools like oe-pkgdata-util help narrow things down, and you can then look in specific work directories22:08
smurrayand no IDE here, find and grep ;)22:08
splatchthat's correct find -exec grep will become my new friend ;-)22:10
*** goliath <goliath!> has quit IRC22:14
*** samvlewis <samvlewis!~samvlewis@> has quit IRC22:15
*** samvlewis <samvlewis!~samvlewis@> has joined #yocto22:16
*** BloodSurfer <BloodSurfer!~BloodSurf@2a02:8070:181:d000:f145:e69a:21b5:90d5> has quit IRC22:23
splatchsmurray: given that you made 100% score in answering my problems, I will try to pull one more thing. The qemu. Builder I have does everything in docker. I have all images and I can run them separatelly (mentioned iso in vbox), but I can't make use of runqemu yet. I see many different layouts of yocto projects - some with oe-init-buildenv script at the top, some with all recipes, and some trimmed down just22:25
splatchto supplied layers. What is the good way to avoid host contamination and preserve reasonable dev velocity?22:25
smurraywhen you say "host contamination", what do you mea, exactly?  I ask, because it usually has a specific meaning in OE / Yocto22:27
smurraysource tree contamination, you mean?22:28
splatchI just repeated smart term I heard in one of videos :)22:28
splatchso the docker had an advantage for me of not bothering to setup the project and have reproducible execution path, but it doesn't give luxury of running qemu22:29
smurrayusually in OE build context it means things from the host machine that you don't want getting picked up in the build process22:29
splatchso that part for me is solved by docker sandbox22:30
smurrayah, you mean being isolated from the host machine wrt build requirements, i.e. having a known working env22:31
splatchyes, my build is made in safe way, so the natural next step is a test run to see if services go up22:31
splatchsince I got qemu build working I am in a need of exiting the sandboxed docker and making runqemu on my host22:32
smurrayI guess your issue is needing sudo access for the network config runqemu wants to do?22:32
smurrayor X access to display?22:33
splatchI'm fine with shell as image I have doesn't ship any X related stuff22:34
splatchI see few output images and few ways telling runqemu script to run them, yet I am not entirely clear which one I should use22:34
smurrayhrm, I'd have thought that you could potentially make runqemu work in docker, then, with e.g. something like "runqemu kvm nographic slirp"22:35
splatchI have ext4, iso, qcow2. I am in process of getting oe-init script on the host to see if it will behave better than within docker where it simply dies22:35
smurraythere are definitely folks who are the channel who do a lot more with docker than I do22:36
splatchit did launch but I get either no output at all in the shell window or no bootable media or other error coming from runqemu22:36
splatchfrom documentation samples I saw that wic image should go if I give runqemu full path, will it?22:37
smurrayI've been sandboxing with the schroot utility locally as it works for my usecases, and I do sometimes use runqemu inside the sandbox, but the default TUN networking does need workarounds wrt sudo22:37
smurraysplatch: yes, that's one option22:38
smurraysplatch: runqemu prints the qemu command line it's running, so cut and pasting that and running it by hand inside the OE env is also an option for debugging22:39
smurraysplatch: I suspect you'll get better answers on the weekend or Monday, pretty quiet here today with the holiday in the US22:40
splatchsmurray: if US will woke up after quarantinee ;-)22:41
smurraysplatch: heh, most folks working on OE/YP work from home22:42
splatchwe had some holidays down the path and plenty of people here took chance just to leave city and be out as long as possible22:42
splatchsmurray: can you tell a bit more about schroot? I have some experiences with chroot recovering my manjaros poor f2fs ;-)22:43
smurraysplatch: it's available on most distros, don't see a homepage for it right off.  man page is here
smurraysplatch: you basically give it a directory with a rootfs and it bind mounts your filesystems into the sandbox env it creats22:46
splatch"schroot does not virtualise the entire system; it only virtualises the filesystem, and some parts of the filesystem may still be shared with the host. It is therefore fast, lightweight and flexible." I like the idea22:47
smurraysplatch: typical usage is to use debootstrap to make a clean Debian or Ubuntu to use22:47
splatchso its another variant of chroot22:47
smurrayyeah, it's a wrapper around it, for the most part.  A lot more convenient to use, though22:47
smurrayit has some support for fancier things wrt the filesystem mounting I don't use22:48
smurraythe big advantage is it's simple22:48
splatchthat's something which is interesting for me as I would like to play with firstboot & data partition22:49
smurrayif you catch some of the other docker users here, they can likely give some tips for using it22:50
smurraysplatch: moto-timo and JPEW are probably the people to ask if they happen to be around22:51
splatchI'm close getting qemu locally, lets see how it will fail :)22:51
smurraygood luck22:51
splatchthanks, and thank you very much for assistance once again!22:52
smurrayno worries22:52
*** JaMa <JaMa!~martin@> has joined #yocto22:53
splatchI had to tune generated bblayers it starts but ends quickly with "FileExistsError: [Errno 17] File exists: '/usr/bin/core_perl/pod2man' -> '/home/splatch/projects/my-bsp/build/tmp/hosttools/pod2man'"23:03
splatchhm. maybe I should have spare checkout of poky somewhere else23:04
*** sgw2 <sgw2!> has quit IRC23:12
*** sgw2 <sgw2!> has joined #yocto23:13

Generated by 2.17.2 by Marius Gedminas - find it at!