**** BEGIN LOGGING AT Mon Feb 10 02:59:58 2020
Feb 10 04:22:23 <yocti> New news from stackoverflow: Build gcc-cross failed, nothing provides 'gcc-cross' <https://stackoverflow.com/questions/60143964/build-gcc-cross-failed-nothing-provides-gcc-cross>
Feb 10 04:52:28 <yocti> New news from stackoverflow: YOCTO:Where does Beaglebone black firmware directory located at core image sato build? <https://stackoverflow.com/questions/60144209/yoctowhere-does-beaglebone-black-firmware-directory-located-at-core-image-sato>
Feb 10 06:52:50 <yocti> New news from stackoverflow: Yocto - File is not present in my image (but present in tmp/work directory in build environment) <https://stackoverflow.com/questions/60145375/yocto-file-is-not-present-in-my-image-but-present-in-tmp-work-directory-in-bu>
Feb 10 07:42:14 <mckoan> good morning
Feb 10 08:53:08 <yocti> New news from stackoverflow: Best way to upgrade Yocto Version <https://stackoverflow.com/questions/60146768/best-way-to-upgrade-yocto-version> || Change keyboard layout on Linux without X11 <https://stackoverflow.com/questions/60116012/change-keyboard-layout-on-linux-without-x11>
Feb 10 09:06:56 <Ad0> is there a variable or something for /proc/sys/net/ipv4/ip_forward ?
Feb 10 09:07:02 <Ad0> setting those kind of settings
Feb 10 09:07:51 <Ad0> https://git.yoctoproject.org/cgit.cgi/poky/plain/meta/recipes-extended/procps/procps/sysctl.conf
Feb 10 10:04:06 <ThomasD13> Hello, I'm debugging the do_configure task of virtual/kernel recipe. In ../temp/run.do_configure I can see what the build server executes. But what does create/defines the run.do_configure script?
Feb 10 11:20:26 <tolszak> Hello, I have a case, my custom package A depends on gsl package, it is added to DEPENDS, but when I populate sdk for image that contains package A the gsl-dev is not installed. How to define A so that its build dependencies are added when sdk is populated?
Feb 10 11:28:10 <d_s_e> Hi, is is possible to output the current revision for svn-based layers in the build configuration similar to git-based layers? for git-layers it shows the current branch and commit-hash, for svn only "<unknown>:<unknown>".
Feb 10 11:29:02 <LetoThe2nd> d_s_e: not by default, AFAIK
Feb 10 11:29:28 <LetoThe2nd> d_s_e: but its only software :)
Feb 10 11:41:39 <d_s_e> LetoThe2nd: I was hoping that somebody has done this already, so I could save the time:-)
Feb 10 11:42:14 <LetoThe2nd> d_s_e: sorry to say, but i guess most rather investerd their time in doing the svn->git transition.
Feb 10 11:48:21 <RP> d_s_e: http://git.yoctoproject.org/cgit.cgi/poky/tree/meta/classes/metadata_scm.bbclass?id=f702509017c11f02c2c89ece4890ae5e9fbbbe84 - setting METADATA_REVISION to the svn function may work
Feb 10 11:49:07 <RP> "This only works with older subversion. For newer versions this function will need to be fixed by someone interested"
Feb 10 11:51:02 <d_s_e> LetoThe2nd: I would prefer do do that, too, but my customer uses svn and won't change this. :-/
Feb 10 11:51:11 <d_s_e> RP: thnaks, I will have a look at this.
Feb 10 11:51:12 <RP> ThomasD13: bitbake constructs those functions based on the variables set by base.bbclass, and for example autotools.bbclass for autotools recipes
Feb 10 11:51:54 <RP> d_s_e: most people using svn moved to git a while ago
Feb 10 11:52:05 <LetoThe2nd> d_s_e: make $YOURCUSTOMER pay for you to be "someone interested"
Feb 10 12:00:12 <d_s_e> LetoThe2nd: that's why I'm hesitant in putting to much effort into this topic. I guess, they would pay, but this feature currently has a very low priority.
Feb 10 12:10:04 <ThomasD13> thanks RP
Feb 10 12:10:50 <LetoThe2nd> RP: can i ask you about a kind of brainfart i had this morning?
Feb 10 12:11:39 <LetoThe2nd> RP: namely - what do you think of dynamically creating recipe data?
Feb 10 12:13:52 <LetoThe2nd> RP: i mean, from my understanding bitbake takes all those text files and parses them into some kind of in-memory storage. could this in-memory-storage be pre-filled with data, therefore being able to skip the textfile on disk?
Feb 10 12:15:26 <RP> LetoThe2nd: that is what our cache is!
Feb 10 12:16:30 <LetoThe2nd> RP: yeah but the cache gets invalidated if the textfile does not match? i'm thinking about inject data there directly, via some form of API
Feb 10 12:17:15 <RP> LetoThe2nd: how would that be more useful that what we have today? Surely you want it to reparse if the textfile doesn't match?
Feb 10 12:17:29 <LetoThe2nd> RP: nah. i don't want a textfile.
Feb 10 12:17:42 <LetoThe2nd> RP: its really just in the brainfart stage at the moment.
Feb 10 12:18:03 <RP> LetoThe2nd: you could inject data quite happily. You then become responsible for its validity though
Feb 10 12:18:13 <LetoThe2nd> thats obvious.
Feb 10 12:18:39 <RP> I just don't understand where the win would be
Feb 10 12:19:05 <LetoThe2nd> Me neither.
Feb 10 12:19:38 <LetoThe2nd> Seriously, its more like a "What if"-line of thought.
Feb 10 12:45:09 <tolszak> Could you please point me some docs how to fix a recipy so that its build dependencies are installed in sdk?
Feb 10 12:45:27 <tolszak> For some reason dev version of package added to DEPENDS is not installed in sdk
Feb 10 12:46:14 <tolszak> I have image that contains package A, A depends on B, and when I populate_sdk for this image package B-dev is not in sdk
Feb 10 12:48:51 <LetoThe2nd> tolszak: technically, if A is already built and can be used in the sdk, you don't need the headers of B
Feb 10 12:50:16 <LetoThe2nd> so this sounds more like you explicitly want B-dev in the recipe, regardless of A being there or not?
Feb 10 12:52:42 <tolszak> LetoThe2nd: I want to generate SDK to build A from 3rd party sdk (Qt Creator). How to ensure that A build deps are in sdk
Feb 10 12:53:51 <LetoThe2nd> tolszak: i'd guess that the image that you do the sdk for then needs to expand TOOLCHAIN_TASK_HOST
Feb 10 12:55:42 <tolszak> LetoThe2nd: I see - thank you
Feb 10 13:32:46 <rburton> LetoThe2nd: if you're doing comment in source code you *need* to cover the boost.bb comments
Feb 10 13:33:16 <LetoThe2nd> rburton: pointer?
Feb 10 13:33:31 <rburton> boost.bb in oe-core, scroll down to the big comment
Feb 10 13:34:37 <LetoThe2nd> rburton: toot long for a tweet.
Feb 10 13:37:08 <LetoThe2nd> rburton: https://twitter.com/TheYoctoJester/status/1226862556972756994
Feb 10 13:38:23 <moustafa> hello everybody
Feb 10 13:38:55 <rburton> LetoThe2nd: :)
Feb 10 13:39:06 <moustafa> I am trying to run some qt file and I tar it with tar cfj file name.tar.xv file name
Feb 10 13:39:43 <moustafa> and when I bitbake it gives me this error => ERROR: dltc-1.0-r0 do_unpack: Unpack failure for URL:
Feb 10 13:48:20 <paulbarker> LetoThe2nd: Nothing beats the "Fuck me harder with a chainsaw" comment in an old version of Linux
Feb 10 13:48:42 <LetoThe2nd> paulbarker: blergh
Feb 10 13:49:11 <LetoThe2nd> i've just had lunch. that was inappropriate.
Feb 10 13:49:45 <LetoThe2nd> (given the side note that its not too long since i've watched Se7en)
Feb 10 13:52:43 <paulbarker> LetoThe2nd: Or this personal favourite of mine: https://github.com/pratyakshs/GeekOS/blob/master/GeekOS/src/geekos/util.asm#L25
Feb 10 13:53:06 <LetoThe2nd> hrhrhr
Feb 10 13:53:12 <paulbarker> It took me much more than 1hr to figure out how those 2 or 3 assembly instructions worked
Feb 10 13:53:32 <LetoThe2nd> paulbarker: now i need to listen to steel panther to recover, and it's your fault!
Feb 10 14:14:09 <JPEW> \
Feb 10 14:15:27 <moustafa> anyhelp ?
Feb 10 14:15:32 <moustafa> I am trying to run some qt file and I tar it with tar cfj file name.tar.xv file name
Feb 10 14:15:36 <moustafa> and when I bitbake it gives me this error => ERROR: dltc-1.0-r0 do_unpack: Unpack failure for URL:
Feb 10 14:17:38 <LetoThe2nd> moustafa: we've seen it, but the question is lacking detail so probably nobody feels like stepping up.
Feb 10 14:18:05 <erbo> moustafa: I think you need to describe exactly what you are doing in more detail, maybe put the recipe you're working on in a pastebin etc.
Feb 10 14:18:25 <erbo> Ah, I'm too slow
Feb 10 14:18:26 <moustafa> aha ok
Feb 10 14:19:21 <moustafa> I am building custom image and trying to type the name of the app to run it
Feb 10 14:20:29 <moustafa> i read about some steps by tar the folder and build beside it <folder name>_1.0.bb file to describe its root
Feb 10 14:20:38 <moustafa> it was working well
Feb 10 14:21:01 <moustafa> but now when I bitbake the file it gives me that error
Feb 10 14:21:25 <neverpanic> moustafa: https://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html#var-FILESPATH
Feb 10 14:21:29 <moustafa> ERROR: dltc-1.0-r0 do_unpack: Unpack failure for URL:........
Feb 10 14:22:02 <moustafa> neverpanic ok I'll check it thanks
Feb 10 14:22:45 <neverpanic> The default value does not include the directory that contains the recipe. However, it does include a files/ directory, a ${PN}/ and a ${BPN}/ directory next to the recipe, so maybe you want to place your tarball there.
Feb 10 14:23:39 <moustafa> aha ok I'll check it
Feb 10 18:52:45 <d_thomas> I would like to set VOLATILE_LOG_DIR="no".  I first added this setting to my rootfs recipe, but while dumping the variables showed the setting was changed there was no effect.  I moved the setting to conf/layer.conf file and I achieved the desired effect.  I'm confused why that is the case.  Did I place this setting in the wrong recipe file or does
Feb 10 18:52:46 <d_thomas> it need to be in a conf file?
Feb 10 19:08:59 <psrcode> Anybody using qemu ppc64 with yocto? I found the layer from Armin (akuster) and adapted it against yocto warrior but I cannot get the kernel to boot, it gets to #Linux... and the crash without much information ... and i'm not that familiar with debugging qemu failure
Feb 10 19:10:49 <khem> d_thomas: set VOLATILE_LOG_DIR = "no" in local.conf
Feb 10 19:12:46 <khem> psrcode: yes I have it working for ppc64le here https://github.com/YoeDistro/openembedded-core/commits/yoe/mut
Feb 10 19:13:24 <psrcode> nice, do you think using the big endian would be a lot of workÉ
Feb 10 19:13:40 <khem> psrcode: especially 2 commits https://github.com/YoeDistro/openembedded-core/commit/9903c4e316475e326f50905914571426a4bcdb61 and https://github.com/YoeDistro/openembedded-core/commit/4033d731e72e7f5fbe0adcea656d7ee3c47875ba
Feb 10 19:14:10 <d_thomas> khem:  Thank you.  Why the conf file instead of a recipe?  Is it just the scoping of settings in recipes?
Feb 10 19:15:04 <psrcode> khem: thanks a lot! I'll try it for sure
Feb 10 20:30:49 <khem> psrcode: are you looking for BE or LE ?
Feb 10 20:43:13 <psrcode> BE
Feb 10 20:43:30 <psrcode> khem: ^
Feb 10 20:45:02 <psrcode> and ideally it would be a layer on top of base oe-core but this is work that can be done on our side later on. We are mostly interested in this to ensure that lttng works in weird scenario were userland and kernel are differecent (64 bit/32 bit)
Feb 10 20:45:12 <psrcode> an correclty on ppc64
Feb 10 20:45:38 <psrcode> we know userspace is okai since we do have worker with old ubuntu on it but still yocto would be a goot fit for our testing infra internally
Feb 10 20:49:50 <winotu> o/  Hi
Feb 10 20:51:38 <winotu> I'm trying to find why some dependecy is not causing rebuilding dependent components in my image. bitbake -g shows proper dependecies. Issue occurs when I change preffered version variable.
Feb 10 20:53:10 <winotu> it looks like some bitbake runtime race condition in checking which version of recipe use and based on that use object from sstate cache
Feb 10 20:56:32 <winotu> can I find some data in workspace how bitbake resolved which recipe it should use for which component/recipe?
Feb 10 20:57:37 <khem> psrcode: yeah it works with BE too, thats what I got going first but ppc64le is more of a thing
Feb 10 20:58:13 <psrcode> khem nice
Feb 10 20:58:23 <khem> you might have to create new defconfig or perhaps just add a .cfg with BE kernel config options and add it to SRC_URI
Feb 10 20:58:55 <psrcode> k, i'll get LE working first and them work my way up to BE
Feb 10 20:59:42 <khem> yeah
Feb 10 20:59:50 <psrcode> thanks a lot again!
Feb 10 21:00:43 <armpit> khem, we most likely need BE machine to include a BE kernel fragment
Feb 10 21:01:14 <armpit> there is a bug open for arm64 regarding the same issue
Feb 10 21:03:11 <armpit> wonder if those kind of generic machines should sit in core
Feb 10 21:06:19 <roussinm> I would like to create a virtual package, but one of the package is in poky, and the other one in meta-oe. I guess it doesn't matter right? That's what virtual is for?
Feb 10 21:07:52 <roussinm> I think it's going to break if people doesn't set the preferred_provider, so I should set a default to not break current builds. Where would I write such default provider? piko.conf? Maybe if there is only one provider it just works?
Feb 11 00:11:24 <khem> armpit: perhaps Bruce should add these defconfigs to kmeta I have already sent him LE ones
Feb 11 00:21:56 <mischief> why does multiconfig need different tmpdirs?
Feb 11 00:22:46 <khem> canonical triplets could conflict
Feb 11 00:24:14 <mischief> khem: what if i only have one triplet and one distro but multiple machines and images?
Feb 11 00:33:55 <khem> then you dont need multiconfig
Feb 11 00:35:08 <khem> well sort of you do, if your final image is collection of images built for different machines
Feb 11 00:48:45 <mischief> no, i don't need that. maybe i misunderstood multiconfig.
Feb 11 00:49:07 <mischief> i wanted to see if the build could be made faster by building for multiple MACHINE at once with a multiconfig.
Feb 11 00:49:29 <mischief> we don't need to make a final image which is a collection of images
Feb 11 00:49:51 <mischief> just N machines * M images
Feb 11 02:12:03 <armpit> khem, you sent Bruce kernel-cache changes for Endiness options ?
Feb 11 02:45:35 <khem> armpit: no, I sent the defconfig that I am carrying he said he would do the needed to turn it into linux-yocto tooling
**** ENDING LOGGING AT Tue Feb 11 02:59:57 2020