**** BEGIN LOGGING AT Tue Apr 30 02:59:57 2019
Apr 30 04:00:27 <armpit> the dependency is not the test but the python script running, most likely need to update hosttools
Apr 30 06:11:11 <yocti> New news from stackoverflow: What is a kernel image? Why is it mandatory in Embedded? <https://stackoverflow.com/questions/55898349/what-is-a-kernel-image-why-is-it-mandatory-in-embedded>
Apr 30 06:41:16 <yocti> New news from stackoverflow: What are the main Things in BB files and Give a Basic Example? <https://stackoverflow.com/questions/55915207/what-are-the-main-things-in-bb-files-and-give-a-basic-example> || Is there a way to create hash(sha256) for images in Yocto's swupdate layer? <https://stackoverflow.com/questions/55915105/is-there-a-way-to-create-hashsha256-for-images-in-yoctos-swupdate-layer>
Apr 30 07:27:37 <mckoan> good morning
Apr 30 08:30:07 <JoeR> Hello Yocto people. I have a simple question that for some reason eludes me.
Apr 30 08:31:08 <LetoThe2nd> JoeR: state your case of emergency and wait for rescue :)
Apr 30 08:31:56 <JoeR> :-) So. I'm writing a recipe for a library. The library is based on cmake, so after inheriting cmake most of the work is done for me.
Apr 30 08:32:18 <JoeR> However. I'm struggling to pass the sysroot to cmake for a particular include path.
Apr 30 08:32:34 <JoeR> It's correct everywhere, apart from the one addition I have to make.
Apr 30 08:32:51 <LetoThe2nd> JoeR: you mean for some dependency, or what?
Apr 30 08:32:56 <JoeR> Yeah.
Apr 30 08:33:21 <LetoThe2nd> JoeR: and you have set the dependency in DEPENDS? so it properly is prepared?
Apr 30 08:33:41 <JoeR> Yup. The file is there. If I pass the entire sysroot path, hard coded (i.e. my home dir etc.) it works.
Apr 30 08:33:48 <JoeR> So I'm definitely passing it to cmake.
Apr 30 08:34:00 <JoeR> I've tried referencing ${sysroot} but that's empty.
Apr 30 08:34:30 <JoeR> And I've tried using the "=/usr/include/somedep" to get the sysroot in that way,. i.e. -I=/usr/include/somdep
Apr 30 08:34:37 <JoeR> But that also didn't work.
Apr 30 08:35:09 <JoeR> So I guess I'm confused, cmake has had it sysroot set, so when my path turns into /usr/include/somedep, I expected that to work. But it doesn't.
Apr 30 08:35:15 <LetoThe2nd> well for one, you have to go through EXTRA_OECMAKE, but i guess you already knew that
Apr 30 08:35:23 <JoeR> Yeah, that's where I'm setting it.
Apr 30 08:35:41 <JoeR> I don't understand why ${sysroot} is empty once it's been passed that way.
Apr 30 08:35:58 <LetoThe2nd> for second, i'd say, inspect the output of bitbake -e YOURRECIPE to find the variable where the sysroot is located
Apr 30 08:36:00 <JoeR> I'm on Krogoth, which I know is ancient, but that's on my list to fix in the future....
Apr 30 08:36:18 <JoeR> I thought the sys root was always in ${sysroot}?
Apr 30 08:36:39 <LetoThe2nd> i don't think so, actually
Apr 30 08:36:49 <JoeR> So that's probably mistake A then :-)
Apr 30 08:37:20 <JoeR> I wonder why when the sysroot does appear to be set correctly rather than in the include path it's not getting searched though.
Apr 30 08:37:51 <LetoThe2nd> JoeR: things you can rely on are set in meta/conf/bitbake.conf
Apr 30 08:38:03 <LetoThe2nd> JoeR: and ${sysroot} is not there.
Apr 30 08:38:15 <JoeR> Thanks, that's worth knowing.
Apr 30 08:39:20 <LetoThe2nd> JoeR: for a current release I think it should be RECIPE_SYSROOT
Apr 30 08:39:40 <LetoThe2nd> JoeR: but for krogoth, please inspect bitbake.conf yourself, i don't have that handy right now.
Apr 30 08:39:46 <JoeR> Will do.
Apr 30 08:41:25 <JoeR> Looks like STAGING_DIR_TARGET, does that sound correct?
Apr 30 08:42:23 <LetoThe2nd> nope
Apr 30 08:42:34 <LetoThe2nd> thats the sysroot it stages *to*
Apr 30 08:42:40 <JoeR> OK.
Apr 30 08:43:31 <JoeR> Hmm. Nothing else obvious in there.
Apr 30 08:44:11 <LetoThe2nd> (its pretty much out of my area of expertise, so i can't really help much more, sorry)
Apr 30 08:44:37 <JoeR> That's ok. Thanks for the pointers! May your spice always flow.
Apr 30 08:47:02 * LetoThe2nd nods distinguishedly and awards 5 grams to JoeR.
Apr 30 08:48:38 <hoxy> hi everyone. I am having a problem where my kernel device tree is not installed to the /boot folder in my rootfs. It is compiled correctly and available in the deploy folder, but I just can't get it to install to my rootfs
Apr 30 08:49:11 <hoxy> I tried writing a bbappend for my kernel image, but it does not work
Apr 30 08:49:11 <hoxy> https://pastebin.com/0cVpt9E2
Apr 30 08:55:32 <LetoThe2nd> hoxy: just add kernel-devicetree to your IMAGE_INSTALL
Apr 30 08:57:01 <LetoThe2nd> hoxy: together with properly setting KERNEL_DEVICETREE in your kernel recipe, this does the whole magic for you.
Apr 30 09:02:03 <hoxy> I already have "IMAGE_INSTALL_append = " kernel-devicetree"" in my image recipe
Apr 30 09:02:16 <hoxy> I forgot to mention that sorry
Apr 30 09:04:23 <hoxy> and the KERNEL_DEVICETREE variable contains my device tree variable
Apr 30 09:04:55 <hoxy> KERNEL_DEVICETREE = "renesas/r8a7795-es1-salvator-x.dtb renesas/r8a7795-salvator-x.dtb renesas/r8a7795-salvator-xs.dtb renesas/r8a7795-salvator-xs-2x2g.dtb renesas/r8a7795-salvator-xs-4x2g.dtb"
Apr 30 09:11:41 <yocti> New news from stackoverflow: Shared libraries cannot be found when building application in Yocto <https://stackoverflow.com/questions/55917197/shared-libraries-cannot-be-found-when-building-application-in-yocto>
Apr 30 09:13:36 <LetoThe2nd> hoxy: well then inspect the contents of the kernel-devicetree package
Apr 30 09:33:35 <hoxy> LetoThe2nd: Okay I looked. The files are in getting installed to the image/boot folder in the package
Apr 30 09:34:41 <LetoThe2nd> hoxy: which is what you wanted, right?
Apr 30 09:35:07 <hoxy> yeah, but when I unpack my rootfs they are not there
Apr 30 09:35:57 <LetoThe2nd> hoxy: is your rootfs doing magic with a separate boot partition that gets mounted somewhere? and does the package really get installed? the manifest will tell you
Apr 30 09:39:39 <hoxy> I don't thing there is a seperate boot partition. Also the kernel image is installed fine.
Apr 30 09:39:52 <hoxy> I'll see whats in the manifest thanks
Apr 30 09:40:12 <LetoThe2nd> hoxy: well if you say that the package contains the files AND that it is installed, then the files are most probably there.
Apr 30 09:44:52 <hoxy> yes kernel_devicetree is in the manifest
Apr 30 09:45:14 <hoxy> I'll just clean the whole thing. Maybe that helps
Apr 30 09:55:07 <hoxy> oh I unpacked an old version of the RFS. Fuck I'm stupid.
Apr 30 09:55:14 <hoxy> Thanks for the help though
Apr 30 09:55:21 <JoeR> lol We've ALL been there.
Apr 30 09:55:33 <hoxy> Helped me to understand how to debug those issues in the future
Apr 30 09:56:20 <LetoThe2nd> hrhrhr
Apr 30 10:30:21 <kanavin> RP: yeah I have been following the discussion, glad the issue is sorted (the kernel tcp hang)
Apr 30 10:30:43 <kanavin> rburton_, sure :)
Apr 30 10:54:50 <RP> kanavin: yes, its a nice success story for our testing too
Apr 30 11:08:28 <RP> armpit: figured out the ltp thing, its a free space issue. Fix pushed
Apr 30 11:12:12 <kanavin> RP: I am unable to reproduce the python3-pycairo failure seen on the AB. I will send the update again, and if the failure again occurs, would be good to get a snapshot of the work directory for the recipe.
Apr 30 11:16:03 <kanavin> RP: never mind, just saw it :)
Apr 30 11:22:11 <RP> kanavin: ok :)
Apr 30 11:52:24 <kanavin> rburton, you owe me a beer :)
Apr 30 12:12:17 <yocti> New news from stackoverflow: How to setup a machine as a mirror server for Yocto when fetching packages? <https://stackoverflow.com/questions/55625512/how-to-setup-a-machine-as-a-mirror-server-for-yocto-when-fetching-packages>
Apr 30 12:13:05 <paulbarker> Looking at the latest systemd release notes, getty@tty1.service is no longer enabled by default. Is this something we should re-enable in oe-core or should we move to fixing it in BSPs?
Apr 30 12:13:59 <paulbarker> For boards with HDMI or other graphics output (Raspberry Pi, BeagleBone Black, etc) we probably want a console on tty1 by default
Apr 30 12:15:12 <paulbarker> I can't see an easy way to re-enable this. It doesn't belong in SERIAL_CONSOLES as it has no baud rate
Apr 30 12:30:06 <yann> When trying to build mesa on a wayland distro, it complains about missing wayland-egl-backend, I can't find a recipe for that, am I missing something ?
Apr 30 12:36:31 <rburton> kanavin: cheers :)
Apr 30 12:36:49 <rburton> yann: your bsp probably ships a custom GL?
Apr 30 12:37:08 <kanavin> rburton, I still intend to do that not-London UK trip :)
Apr 30 12:38:49 <RP> kanavin: what did rburton bribe you with?!
Apr 30 12:38:57 <rburton> "money"
Apr 30 12:39:09 <rburton> i have a US $2 reserved for kanavin
Apr 30 12:40:10 <erakis> Hi, I'm not sure what I did but before pulling new change of fido branch for bitbake and some meta, I had ipk package like "busybox_1.23.1-r0_cortexa8hf-vfp-neon.ipk", now I'm getting "busybox_1.23.1-r0.0_cortexa8hf-vfp-neon.ipk". What is the .0 added ? Where it comes from ? This is annoying because now `opkg list-updatable` list out all the packages instead of just the ones that have changed.
Apr 30 12:41:17 <rburton> prservice adds those
Apr 30 12:41:29 <yann> rburton: that's me-amd#sumo, shipping a customized mesa - hm, I see that official mesa-17 package indeed *builds* libwayland-egl, whereas the mesa-19 package there requires it - looks like it would have been split (and I've already noticed how meta-amd breaks wayland support in several places
Apr 30 12:41:47 <yann> read "that's meta-amd#sumo" :)
Apr 30 12:47:26 <erakis> @rburton: That make sens. I comment #INHERIT += "buildhistory"    #BUILDHISTORY_COMMIT = "1"   #PRSERV_HOST = "localhost:0", run cleansstate taks on the image and then "rm -rf ./tmp". Hope it will fix it ? I needed the PR service, but I think I'll have to do without it.
Apr 30 12:50:28 <erakis> rburton: I'm still getting the ".0". Is there something else I can do ? Maybe once the PR service is enabled, there is not enough to comment lines to disable it ?
Apr 30 13:42:33 <yocti> New news from stackoverflow: How to tell bitbake to calculate variable's basehash value after specific task? <https://stackoverflow.com/questions/55911423/how-to-tell-bitbake-to-calculate-variables-basehash-value-after-specific-task>
Apr 30 13:50:57 <armpit> RP, k..
Apr 30 14:43:56 <RP> LetoThe2nd: somehow we've managed to make the twitch stream conflict with the engineering sync call :(
Apr 30 14:44:09 <RP> I thought the times were different but they're not
Apr 30 14:50:11 <LetoThe2nd> RP: no worries, we'll certainly manage
Apr 30 14:50:22 <LetoThe2nd> ndec: seems to be online already too
Apr 30 14:50:47 <RP> LetoThe2nd: ok. I'll try and get the other meeting done as quickly as possible
Apr 30 14:51:00 <ndec> well.. i messed up with CET too ;) i thought it was at 3PM for me!
Apr 30 14:51:28 <LetoThe2nd> i triple checked and even looked with vim into the ics file
Apr 30 14:51:42 <LetoThe2nd> which says 17:00 Europe/Berlin
Apr 30 14:52:19 <ndec> hmm. i think 5 CET is 6PM these days, since DST.
Apr 30 14:52:47 <LetoThe2nd> thats why i checked, but it explicitly says Eure/Berlin
Apr 30 14:53:07 <LetoThe2nd> and i'm really sure about the berlin time right now :)
Apr 30 14:53:25 <ndec> i know the French time now, since I need to pick my kid at school ;)
Apr 30 14:53:33 <LetoThe2nd> hehe
Apr 30 14:53:34 <JPEW> Err, whats the UTC time?
Apr 30 14:53:43 <ndec> so, you are saying it's in 10 mins?
Apr 30 14:54:17 <LetoThe2nd> ndec: basically. if everything goes nuts, we can do a little longer and show people arriving "late" something too
Apr 30 14:54:38 <ndec> ok.. i will join with a couple of mins late..
Apr 30 14:54:39 <RP> JPEW: UTC, GMT, BST or CET? :)
Apr 30 14:54:56 <JPEW> CDT if you want to make it easy for me :)
Apr 30 14:55:11 <LetoThe2nd> JPEW: C Development Toolkit a.k.a. Eclipse?
Apr 30 14:55:13 <RP> JPEW: easy answer "its starting in 5 mins"
Apr 30 14:55:20 <JPEW> Right
Apr 30 14:55:51 <JPEW> Hmm, apparently our IT department thinks people shouldn't be watching twitch while at work
Apr 30 14:55:53 * armpit 8am PDT, 11am EDT ?
Apr 30 14:57:08 <LetoThe2nd> we certainly need a better way
Apr 30 14:57:19 <LetoThe2nd> next time, we will start at 6:66 UGT
Apr 30 14:57:22 <armpit> 2 cans and a string?
Apr 30 14:57:35 <LetoThe2nd> which is both metal and very un-ambiguous
Apr 30 15:00:13 <LetoThe2nd> so, my clock says 5pm
Apr 30 15:00:26 <armpit> miller time ?
Apr 30 15:00:41 <LetoThe2nd> armpit: its always beer o'clock somewhere
Apr 30 15:01:21 <Crofton>  rul?
Apr 30 15:01:35 <zeddii> https://www.twitch.tv/yocto_project
Apr 30 15:05:21 <florian> heh
Apr 30 15:06:39 <florian> reading a short tutorial might be faster than whating this guys high quality video ;)
Apr 30 15:12:13 <tlwoerner> different people like to learn in different ways; some people might prefer to read about something, others might prefer to watch (a video of) someone do(ing) it
Apr 30 15:12:53 <yocti> New news from stackoverflow: How to pass sysroot from Yocto to cmake <https://stackoverflow.com/questions/55923455/how-to-pass-sysroot-from-yocto-to-cmake>
Apr 30 15:14:00 <florian> yep... its not the worst idea ever. for someone no idea about all this it might be nice
Apr 30 15:18:46 * armpit maybe twitch has a howto ?
Apr 30 15:19:29 <zeddii> questions in the twitch chat. an intel person should answer :D
Apr 30 15:20:26 <florian> indeed
Apr 30 15:20:50 <florian> in fact it pretty much looks like he did this before some times
Apr 30 15:21:22 * zeddii notes RP has a sword icon. fancy.
Apr 30 15:21:57 * armpit Brits
Apr 30 15:22:01 <RP> zeddii: moderator superpowers!
Apr 30 15:22:44 <zeddii> ahah!
Apr 30 15:22:59 * zeddii sees hovering over the icon would have told him that.
Apr 30 15:23:08 <florian> uh
Apr 30 15:24:26 <Crofton> Who iwll he kick out first
Apr 30 15:26:39 <Crofton> I was wondering how long to get a beer. Next he will play metal
Apr 30 15:27:18 <zeddii> :D
Apr 30 15:27:35 * zeddii approves
Apr 30 15:28:06 <florian> There is no yocto beer so far, right?
Apr 30 15:31:19 <RP> zeddii: ndec gets a crown too
Apr 30 15:31:53 <florian> Hmm... nice, noone complained when he mentioned sysvinit :-)
Apr 30 15:33:24 <zeddii> faaancy indeed!
Apr 30 15:54:13 <LetoThe2nd> phew!
Apr 30 15:54:35 <RP> LetoThe2nd: nicely done! :)
Apr 30 15:55:04 <zeddii> indeed!
Apr 30 15:55:21 <LetoThe2nd> RP: some rough edges still... the browser, and me forgetting to get rid of the old build initially... dang
Apr 30 15:55:41 <LetoThe2nd> plus, i forgot to push the "record" button. hopefully we can download it from twitch
Apr 30 15:56:38 <RP> LetoThe2nd: for a first one it seemed to go well
Apr 30 15:56:52 <florian> LetoThe2nd: not bad indeed... for more credibility you shoud drink coffee or mate :-)
Apr 30 15:57:01 <LetoThe2nd> florian: i am not credible.
Apr 30 15:57:48 <florian> LetoThe2nd: more than most of us I guess :-)
Apr 30 15:59:00 <moemoe> anybody with a short hint what creates "/var/lock/ -> ../run/lock" on boot (thud w/ systemd) - I need a service to depend on it, as it currently fails on boot as the symlink "/var/lock" is not created at execution time.
Apr 30 15:59:13 <ndec> LetoThe2nd: thanks a lot for what you've done... i remember a few weeks ago when you throw out this idea here ;)
Apr 30 15:59:33 <LetoThe2nd> florian: i actually pondered doing it without a beer, but then decided that its worth the fun. if we're all being super serious, then we can also be sales drones doing webinars. which is what i explicitly do not want to do
Apr 30 15:59:37 <moemoe> at least a "ls -l /var" injected into the init script only outputs run, spool, tmp, volatile.
Apr 30 15:59:52 <RP> LetoThe2nd: I'd have had a beer FWIW ;-)
Apr 30 16:00:06 <florian> heh
Apr 30 16:00:11 <LetoThe2nd> ndec: you're welcome :) hopefully i got the line between advertising, explaining, and attracting folks done right.
Apr 30 16:00:44 <LetoThe2nd> so whomever i called RPs minion, please excuse me ;-)
Apr 30 16:00:48 <RP> LetoThe2nd: its needs to have a fun element to it, not too serious
Apr 30 16:01:23 <ndec> LetoThe2nd: are the chat messages going to disappear when I exit? or are they stored anywhere?
Apr 30 16:01:26 <florian> LetoThe2nd: for improvement: yes it really looks like you have done this before which is pretty good... even better would be to have a short series covering a set of common tasks newbies run into. i.e. adding and using layers, creating a derived image or adding a bbappend.
Apr 30 16:01:28 <RP> LetoThe2nd: I should explain that, I just had a really bad experience of a leader once who did refer to "his minions" and I really objected strongly to that attitude
Apr 30 16:02:16 <LetoThe2nd> RP: ah ok.
Apr 30 16:03:20 <LetoThe2nd> florian: thats actually the concept, to iron out a couple core core tasks that beginners will run into every time. my content focus is absolutely on the introductory stuff, and hopefully we can find some others to step up for more deep dive stuff
Apr 30 16:03:21 <florian> LetoThe2nd: and the placement of yourself with the camera facing the window was not the best.
Apr 30 16:03:36 <armpit> RP, pybootchartgui.py needs cairo no mater which format is chosen
Apr 30 16:03:45 <LetoThe2nd> florian: i'll complain to the pub about having windows!
Apr 30 16:03:53 <LetoThe2nd> (pun intented)
Apr 30 16:03:55 <florian> hehe
Apr 30 16:06:06 <LetoThe2nd> and twitch tells me the video is already up for anybody to watch
Apr 30 16:06:38 <LetoThe2nd> ndec: they are going to disappear, and there is no download way unfortunately, please manually copy-paste.
Apr 30 16:07:50 <ndec> LetoThe2nd: thanks. done.
Apr 30 16:09:02 <LetoThe2nd> ndec: great. actually, for reasons unknown we had a lot of viewers (26 svg, 35max), but only a handful people of actually interacting with the chat.
Apr 30 16:09:05 <LetoThe2nd> strange
Apr 30 16:09:27 <ndec> yeah. i noticed that..
Apr 30 16:09:29 <moemoe> okay, my question was a bit unprecise - of course /run/ exists, but not /run/lock as it has to be created inside the tmpfs.
Apr 30 16:10:10 <florian> LetoThe2nd: I'm one of them... too lazy to create myself an account for the chat.
Apr 30 16:10:22 <LetoThe2nd> ndec: an area to improve
Apr 30 16:10:51 <LetoThe2nd> so thats it for me right now, gotta get some food and head home
Apr 30 16:10:59 * LetoThe2nd AFKs
Apr 30 16:11:38 <ndec> florian: ah ok.. you need an account for the chat, otherwise you can 'view' but without the chat. that's what you are saying?
Apr 30 16:12:01 <florian> ndec: right
Apr 30 16:12:10 <ndec> if so, then we should mention that to everyone, so that people can choose deliberately to join (or not) the chat.
Apr 30 16:13:01 <florian> ndec: the ui looks like you can just chat but when you try to send a message it asks you to log in or register first
Apr 30 16:14:37 * florian heading home now... bbiab
Apr 30 16:20:42 <moemoe> ftr: adding "After=systemd-tmpfiles-setup.service" to the systemd unit fixed the problem
Apr 30 16:26:52 <RP> halstead: its looking like we're going to need python3-cairo on the autobuilder workers :/
Apr 30 18:36:17 <aehs29> kanavin: you around?
Apr 30 18:43:40 <halstead> RP, python3 cairo is installed cluster-wide.
Apr 30 18:44:06 <halstead> RP, Except for the perf machines. Waiting until they are idle first.
Apr 30 18:59:59 <RP> halstead: thanks :)
Apr 30 19:00:22 <RP> halstead: very much appreciated!
Apr 30 20:58:57 <RP> armpit: new way for ltp to fail: https://autobuilder.yoctoproject.org/typhoon/#/builders/95/builds/2 :(
Apr 30 20:59:20 <RP> bluelightning: I'm reporting that one to armin here...
Apr 30 20:59:21 * armpit looks
Apr 30 21:14:38 <armpit> RP, do we have any raw qemu logs?
Apr 30 21:15:16 <armpit> I wonder if qemu died
Apr 30 21:19:23 * armpit sched failed or caused the hang
Apr 30 21:21:19 <RP> armpit: probably have to go digging on the autobuilder
Apr 30 21:23:00 <RP> armpit: now I think about it I bet it was out of memory
Apr 30 21:23:15 <RP> armpit: I can try again increasing that, that would explain the ssh connection disappearing
Apr 30 21:24:56 <armpit> I was sure there is a qemu log somewhere but haven't found it
Apr 30 21:25:55 <armpit> being out of memory is now my problem too ; )
Apr 30 21:30:40 <armpit> looks like I had 'IMAGE_ROOTFS_EXTRA_SPACE = "1524288" '
Apr 30 21:31:56 <kergoth> huh, vscode+sshcode works pretty nicely with bitbake projects on remote servers, way nicer than sshfs for sure
Apr 30 21:32:01 <kergoth> might have to play with this some more
Apr 30 21:32:21 <kergoth> course sshcode is just a code-server wrapper, but it's handy
Apr 30 21:33:28 <kergoth> https://www.dropbox.com/s/9wj8sjtx992tk63/Screenshot%202019-04-30%2014.33.21.png?dl=0
Apr 30 21:56:48 <paulbarker> kergoth: Nice! I've been using `rclone mount` to wrap sshfs with some caching but may be worth looking at sshcode
Apr 30 21:57:47 <kergoth> the nice thing is its running on the remote server, not locally, you're just viewing it remotely. but it's still quite responsive, and the I/O is a ton faster without transferring it over ssh
Apr 30 21:59:29 <paulbarker> *bookmarks for later in the week*
Apr 30 22:39:41 <RP> armpit: memory in qemu, not space
Apr 30 22:40:26 <armpit> ah
**** ENDING LOGGING AT Wed May 01 02:59:57 2019