**** BEGIN LOGGING AT Wed Aug 02 03:00:01 2017
Aug 02 07:49:38 <wouterstreamit> Im making a recipe for a program which depends on a library A which in turn depends on another library B. Now my program does not compile when I just depend on library A as a header file in B included by A will be missing. Is it recommended to add B as a dependency myself (which isnt really true imo) or is there some other way to avoid this?
Aug 02 07:56:04 <TafThorne> Sounds like Library A's recipie is incorrect as it does not include all of its own dependencies.
Aug 02 07:56:14 <bluelightning> wouterstreamit: if you need B's header to compile then you really do have a dependency on B
Aug 02 07:57:27 <wouterstreamit> A's recipe is correct and in its recipe-sysroot the correct header files are available, but in its install-sysroot the headers from B are not
Aug 02 07:57:30 <TafThorne> wouterstreamit: is B's header file included anywhere in your program code?
Aug 02 07:57:39 <wouterstreamit> No
Aug 02 07:57:45 <wouterstreamit> Only through a header imported from A
Aug 02 07:59:37 <TafThorne> I'd agree that it is not your dependency but the quick solution would be to include the dependency on B yourself.
Aug 02 08:35:12 <LetoThe2nd> i'm seeing that a buildhistory-enabled image builds fine, but the corresponding esdk build fails because it tries to access a nonexiting sstate-package-sizes.txt
Aug 02 08:35:16 <LetoThe2nd> is this a known issue?
Aug 02 08:36:53 <LetoThe2nd> does this coincide with SDK_INHERIT_BLACKLIST="buildhistory icecc" from populate_sdk_ext bbclass?
Aug 02 09:07:18 <LetoThe2nd> meh. buildhistory was explicitly inherited in local.conf - nevermind
Aug 02 09:32:53 <bluelightning> LetoThe2nd: SDK_INHERIT_BLACKLIST is supposed to disable it if it's enabled
Aug 02 09:34:33 <LetoThe2nd> bluelightning: maybe an evaluation order issue?
Aug 02 09:35:00 <bluelightning> LetoThe2nd: possibly, but SDK_INHERIT_BLACKLIST gets turned into INHERIT_remove
Aug 02 09:35:08 <bluelightning> maybe that no longer works, not sure
Aug 02 09:35:30 <LetoThe2nd> bluelightning: do we have recent changes in that area? project is on krogoth
Aug 02 09:35:46 <bluelightning> no, we did that a while ago, and it did work originally
Aug 02 09:36:14 <LetoThe2nd> bluelightning: hmkay. if i find the time and motivation i'll have a deeper look
Aug 02 10:21:20 <T_UNIX> why would about every task I try to claim "ERROR: Task do_cleansstate is marked nostamp, cannot invalidate this task"
Aug 02 10:21:23 <T_UNIX> ?
Aug 02 10:22:01 <bluelightning> T_UNIX: there's no point using -f with -c cleansstate because it will always run when requested
Aug 02 10:22:25 <T_UNIX> I'm not using -f with it though O.o
Aug 02 10:22:35 <bluelightning> are you using -C then perhaps?
Aug 02 10:22:42 <T_UNIX> yes
Aug 02 10:22:48 <T_UNIX> ah, that's short for -c -f ?
Aug 02 10:22:53 <bluelightning> yes
Aug 02 10:22:56 <bluelightning> well, sort of
Aug 02 10:23:10 <T_UNIX> okay. thanks :)
Aug 02 10:23:48 <bluelightning> actually what -C does is force the specified task but run the default task, so you can say force compilation and have the recipe continue to packaging etc.
Aug 02 10:41:18 <Crofton|work> anyone recall when already stripped QA became an error in poky?
Aug 02 10:44:48 <kanavin> rburton: unfortunately I don't have a specific list of such recipes, and no idea how to produce one (other than remove python 2 from the host and see what breaks)
Aug 02 10:45:51 <rburton> Crofton|work: 2013
Aug 02 10:46:00 <Crofton|work> lol
Aug 02 10:47:03 <rburton> kanavin: yeah we need bb-whatdepends to work again
Aug 02 10:51:42 <kanavin> rburton: but host python 2 is an implicit dependency, it's not written into recipes at all
Aug 02 10:54:58 <rburton> yeah
Aug 02 10:55:13 <rburton> kind of tempted to see what breaks with python removed from hosttools
Aug 02 10:55:55 <LetoThe2nd> rburton: go for it!
Aug 02 10:55:55 <kanavin> rburton: we're just discussing the exact same thing with RP on the internal channel, why don't you join :)
Aug 02 10:56:02 <kanavin> so I don't have to type things twice ;)
Aug 02 10:56:16 <rburton> lol
Aug 02 10:56:34 <Crofton|work> why discuss on internal cabal channel?
Aug 02 10:56:46 <rburton> well, exactly. this is public :)
Aug 02 10:57:02 <kanavin> ask ipuustin :)
Aug 02 10:57:05 <kanavin> he kicked this off
Aug 02 10:59:13 <rburton> Crofton|work: you realise the proper fix would be to fix qwt right
Aug 02 11:00:11 <Crofton|work> Yes, I discussed with bluelightning
Aug 02 11:00:23 <Crofton|work> this is qt4, which is basically life support
Aug 02 11:00:34 <Crofton|work> I checked qt5 recipe and the problem isn't there
Aug 02 11:01:15 <rburton> can i mock gnuradio for still using qt4?
Aug 02 11:01:32 <rburton> kanavin: ca-certs is first on the list of needs a host py2
Aug 02 11:01:43 <LetoThe2nd> möck it! its more evil with möre döts!
Aug 02 11:01:54 <Crofton|work> rburton, please do
Aug 02 11:02:02 <Crofton|work> although there is a qt5 branch
Aug 02 11:02:53 <Crofton|work> I am cleaning up a pull request queue
Aug 02 11:03:04 <Crofton|work> clearly not many people are using qt4 with poky
Aug 02 11:17:43 <peacememories> hi everyone. i'm trying to wrap my head around swupdate right now^^ question: what's the easiest way to modify the u-boot environment for a specific image (or a specific machine) in yocto?
Aug 02 11:18:22 <peacememories> i'm trying to get how to set up my device to have a main system and an update system, and boot to the update system when prompted
Aug 02 11:18:38 <peacememories> (the device is a digilent zybo)
Aug 02 11:25:41 <rburton> kanavin: fwiw, doing the no-python test build in the background now
Aug 02 11:26:07 <hnje> peacememories: boot when prompted? e.g. when you cannot boot the main system?
Aug 02 11:26:17 <kanavin> rburton: the results might be discouraging
Aug 02 11:26:34 <rburton> kanavin: not too bad so far
Aug 02 11:26:42 <peacememories> hnje that (with bootlimit) would be my first try. just trying to get started somewhere right now^^
Aug 02 11:31:13 <hnje> peacememories: with rauc i've a custom u-boot-env for each machine and a uboot-script that gets loaded and handles bootorder/retry
Aug 02 11:46:24 <zepto88> having some trouble with dbus, dbus-send return Failed to open connection to "session" message bus: Could not parse server address: Unknown address type. I works fine from commandline but i want to run this from a c++ system call. Worked fine prior to pyro
Aug 02 13:34:16 <otavio> khem: http://termbin.com/bhiq and http://termbin.com/3hju
Aug 02 14:33:55 <peacememories> can i somehow inherit other image recipes?
Aug 02 14:36:54 <rburton> require <recipe name>
Aug 02 14:44:12 <peacememories> thanks
Aug 02 14:45:54 <peacememories> also, when using swupdate to update my kernel i have to "file" update the uImage and stuff, right? this seems very error prone :/
Aug 02 14:57:04 <TafThorne> Anyone got a Jenkins pipeline job to use repo to run a Yocto build?  That is enough different tools that I am not sure where to ask for help.
Aug 02 15:25:23 <evadeflow> Hi, all. Wondering if anybody can help me work around a sort of chicken-and-egg problem. I need to stuff the contents of a Yocto-generated cpio image into the corresponding ext4 image.
Aug 02 15:25:51 <evadeflow> It's a bit 'weird', and I've been doing it up to now on our Jenkins server, but... it kinda sucks that this one step lives outside of Bitbake.
Aug 02 15:26:09 <evadeflow> This gist is the best way to understand what I'm *trying* to do: https://gist.github.com/evadeflow/d6796747e28d07bb34d18a0b4c13bb5e
Aug 02 15:33:02 <rburton> evadeflow: why not put the cpio into the rootfs before it becomes a ext4
Aug 02 15:33:32 <rburton> oh you're not putting the image inside itself are you?
Aug 02 15:33:51 <rburton> iirc debug2fs can edit a ext4 without mounting it
Aug 02 15:35:21 <evadeflow> Yeah, rburton, I'm putting it inside itself. `:-]
Aug 02 15:35:30 <rburton> rootfsception
Aug 02 15:35:35 <evadeflow> LOL.
Aug 02 15:35:50 <evadeflow> Well, I'm putting the cpio file in the ext4 file.
Aug 02 15:36:39 <rburton> yeah debugfs has  a "write <source> <dest>" operation, so you can write a file inside the file system
Aug 02 15:36:57 <evadeflow> For realz?
Aug 02 15:37:02 <evadeflow> What does that look like?
Aug 02 15:38:09 <evadeflow> I'm not familiar with debugfs—is that a Bitbake thing, or Linux thing?
Aug 02 15:38:29 <evadeflow> https://linux.die.net/man/8/debugfs
Aug 02 15:38:31 <evadeflow> Oh, Linux.
Aug 02 15:38:51 <evadeflow> Thanks for the suggestion, I'll see if I can wire something up!
Aug 02 15:39:24 <peacememories> hmm, can i somehow change the output type of an image to be .ext3.gz (or ext4.gz) in addition to .tar.gz?
Aug 02 15:39:30 <rburton> yeah its part of e2fsprogs
Aug 02 15:39:39 <rburton> peacememories: IMAGE_FSTYPES
Aug 02 15:40:09 <peacememories> rburton thank you, will try
Aug 02 15:43:12 <peacememories> also, apparently `devtool` doesn't find my recipe in workspace/recipes/images/ when trying to use `devtool edit-recipe` :/
Aug 02 15:43:18 <rburton> evadeflow: just verified it works as expected to.  debugfs -w image.ext4; write /some/source/file /where/you/want/in/fs; quit
Aug 02 15:43:34 <rburton> all as non-root
Aug 02 15:43:45 <evadeflow> Oh, man—*thank you*!
Aug 02 15:44:18 <evadeflow> It was gonna take me at least 20 more minutes to finger-mumble my way to that command.
Aug 02 15:44:20 <peacememories> apparently adding -a helps, but i'm not sure why
Aug 02 16:28:50 <Lampel> anyone worked with IMX.6 board ??
Aug 02 16:43:42 <lampel> anyone worked with QWKS-SCM-IMX6DQ board ????
Aug 02 17:10:54 <lsandov1> lampel: better to indicate the issue
Aug 02 19:54:49 <sveinse> What is the fastest method to have a web-browser window (including js and the works) on a framebuffer (non-X11) display device?
Aug 02 20:30:38 <C-o-r-E> what is the fastest way to change global cflags for a bitbake build
Aug 02 20:31:50 <C-o-r-E> specifically when building for the host machine as part of the SDK packages?
Aug 02 20:32:36 <C-o-r-E> i've read that you can put something in a machine conf file
Aug 02 20:32:52 <C-o-r-E> but i only have a machine conf for the target (an arm board)
Aug 02 20:33:17 <C-o-r-E> but the packages want to effect are x86_64
Aug 02 21:23:14 <bbarr> Hey all, has anyone had any success in making a python-scipy package for yocto?
Aug 02 21:28:43 <bbarr> I'm sure I'm not the only person that has asked this.
Aug 02 21:28:48 <bbarr> I couldn't find anything on the net.
Aug 02 21:42:36 <bbarr> Well, I only found evidence that scipy doesn't like being cross compiled.
Aug 02 21:51:58 <flk> bbarr: As far as I know, numpy.distutils has no support for cross compilation
Aug 02 22:48:23 <bbarr> flk: that's what I was afraid of.
Aug 03 00:12:53 <faulteh> hi so is downloads.yoctoproject.org and the mirror at sources.openembedded.org both down? I can't seem to download native shim x86_64-nativesdk-libc.tar.bz2 from either, both locally or a vps in the usa.
Aug 03 02:34:32 <rheagar> hi , does anyone know to to disable prelink function for a specific bb file ? thanks
Aug 03 02:34:55 <rheagar> android has a bool  var "LOCAL_PRELINK_MODULE"  to control this .
Aug 03 02:34:58 <rheagar> thanks
Aug 03 02:41:01 <rheagar> oh i  see i can add that module to blacklist of prelink .
**** ENDING LOGGING AT Thu Aug 03 03:00:01 2017