**** BEGIN LOGGING AT Fri Jul 13 03:00:01 2018
Jul 13 07:22:03 <yocti> New news from stackoverflow: Compulab Yocto: Failure in expanding PATH <https://stackoverflow.com/questions/51319350/compulab-yocto-failure-in-expanding-path>
Jul 13 11:07:44 <tgoodwin> Marex: thanks, will do!
Jul 13 11:54:28 <yocti> New news from stackoverflow: Yocto. Package version generated by git command <https://stackoverflow.com/questions/51324033/yocto-package-version-generated-by-git-command>
Jul 13 12:09:28 <rburton> RP: drop pycairo from next please
Jul 13 12:13:42 <RP> rburton: gone, also removed xmlrunner
Jul 13 12:14:22 <rburton> yay
Jul 13 12:14:34 <rburton> send patch, forward to qa
Jul 13 12:17:42 <rburton> fired a new mut to run over lunch
Jul 13 13:30:40 <yates> if my linux kernel PN = linux-variscite, then is it invalid to have both a linux-variscite_%.bbappend and a linux-variscite_x.y.z.bbappend overrides in my own layer?
Jul 13 13:43:59 <JPEW> yates: No, both will be applied
Jul 13 13:46:00 <yates> is the purpose of the do_preconfigure_prepend() defined in the documentation, as well as other similar "standard" scripts?
Jul 13 13:46:14 <yates> JPEW: ok, thanks.
Jul 13 13:46:40 <yates> of course the one .bbappend will be tied to the version, right?
Jul 13 13:48:22 <JPEW> yates: correct
Jul 13 13:49:29 <yates> what if i had multiple .bbappends that i did not want tied to the version? is the convention to create multiple <meta-mylayer>/recipes-kernel/abc folders, each with a linux_%.bbappend files?
Jul 13 13:49:47 <yates> ..and i wanted to maintain them separately..
Jul 13 13:52:08 <eduardas_m> hello, how can I modify a device tree to get one GPIO pin to change from HI to LO on suspend  (when entering suspend to RAM via "echo mem > /sys/power/state")? The reverse should happen on resuming from suspend.
Jul 13 13:53:03 <eduardas_m> or am I thinking incorrectly about this and GPIO manipulation should be done in userspace via systemd?
Jul 13 13:53:23 <yates> eduardas_m: either way, ##kernel would probably be a better place to ask
Jul 13 13:53:41 <yates> i suspect this wouldn't be a dt configuration, but not sure
Jul 13 13:53:47 <JPEW> yates: You could.... but there are probably better ways than maintaining multiple bbappends like that
Jul 13 13:54:29 <yates> e.g. via includes?
Jul 13 13:54:55 <yates> otherwise, do tell..
Jul 13 13:58:04 <yates> what is the purpose of a do_preconfigure_prepend(). i cannot find it in the docs
Jul 13 14:00:32 <eduardas_m> yates: sorry if my question seemed inappropriate for this channel, it's just that I know there are a lot of people here doing embedded Linux work and having to solve similar problems
Jul 13 14:01:53 <yates> eduardas_m: i don't have a problem with your question at all. it just might benefit you better to ask in the other channel.
Jul 13 14:05:23 <JPEW> yates: It depends on what you are trying to do
Jul 13 14:06:32 <yates> JPEW: i want to have a .bbappend that applies a fixed tweak, but also allow the devtool process to autogenerate the .bbappend and patch files
Jul 13 14:07:31 <yates> is devtool smart and will not touch pieces of the .bbappend file which don't apply to the patches it generated?
Jul 13 14:07:43 <JPEW> yates: Hmm, not sure. I don't use devtool FWIW, so I don't think I can help you there :)
Jul 13 14:07:52 <yates> np
Jul 13 14:07:57 <yates> helps to talk it out in any case.
Jul 13 14:25:04 <yates> how do i refer to the path of a .bbappend file? it appears that THISDIR is the path of the .bb file that is being appended and not the path of the .bbappend file itself.
Jul 13 14:25:49 <yates> i want to define a shell function in that .bbappend that copies a file in a subdirectory of the .bbappend file
Jul 13 14:26:01 <fray> -in a bbappend-
Jul 13 14:26:01 <fray> FILESEXTRAPATHS_prepend := "${THISDIR}/${P}:"
Jul 13 14:26:26 <fray> that would add the path to the BBAPPEND (dir), and a subdirectory of the name of the package to the search path for patches
Jul 13 14:27:00 <fray> See the ':='  causes an immediate evaluation
Jul 13 14:27:06 <fray> (I'm not 100% sure that is necessary)
Jul 13 14:28:36 <yates> fray: +2!
Jul 13 14:28:53 <yates> "=" and ":=" make a difference in this context
Jul 13 14:29:25 <fray> ok.. some variables (very few) are immediately evaluated.. while most others (sounds like THISDIR) is a late evaluation
Jul 13 14:35:25 <yates> does no one know what the do_preconfigure_prepend() function purpose is? i've grepped all files in my project and cannot find where it is invoked, only the only place it is defined.
Jul 13 14:36:17 <yates> my vendor-supplied kernel (variscite) use the function in their kernel recipe: https://paste.fedoraproject.org/paste/dVlRyz8e9Iz8Yoxu8sv-nQ
Jul 13 14:36:50 <yates> it copies the defconfig to ${WORKDIR}/defconfig. that's all well-and-good.
Jul 13 14:36:51 <fray> someone added a task 'do_preconfigure' to a recipe..
Jul 13 14:37:02 <fray> do_preconfigure_prepend then prepends to that task the items included there
Jul 13 14:37:57 <fray> do_preconfigure is not a default task in OE..  so you'll have to find where it's defined.  (It might be part of OE, or the [evil] vendor tree...  but there really is no magic to it)
Jul 13 14:38:09 <fray> you'll see an 'addtask do_preconfigure before ... after ...' somewhere
Jul 13 14:39:44 <yates> most things are "simple" when you understand them and "hard" when you don't...
Jul 13 14:40:06 <yates> yes i've found that in a .bbclass file from fsl
Jul 13 14:40:08 <yates> thanks fray
Jul 13 14:40:24 <yates> galois theory, e.g.
Jul 13 14:41:21 <yates> well, that's not a good example. galois theory torqued my brain even after studying it for a few weeks back in grad school.
Jul 13 14:51:13 <acrap> Hi, folks!
Jul 13 14:51:57 <acrap> I need an encryption support in db, so i've created a patch (bbappend)
Jul 13 14:53:04 <acrap> so, db compiles with encryption support, but these changes broke my image somehow
Jul 13 14:54:02 <acrap> I got https://paste.gnome.org/pz61nyzge
Jul 13 14:55:59 <acrap> I get the error at do_rootfs phase
Jul 13 14:56:35 <acrap> The db recipe https://github.com/openembedded/openembedded/blob/master/recipes/db/db_5.1.19.bb
Jul 13 14:56:55 <acrap> oh, sorry, my version is 5.3.28
Jul 13 14:57:57 <acrap> The right recipe : https://paste.gnome.org/pkyqy9bnf
Jul 13 14:58:45 <acrap> that's my patch: https://paste.gnome.org/pwklgu3zh
Jul 13 14:59:44 <acrap> I've removed --disable-cryptography and addded --with-cryptography
Jul 13 15:14:19 <yates> how is do_task_append() related to d_task()?
Jul 13 15:14:56 <yates> fray mentioned addtask, but that only seems to establish dependencies for do_task
Jul 13 15:16:00 <yates> corretion: how is do_task_prepend() related to do_task()?
Jul 13 15:19:35 <yates> if you define and addtask do_task(), does bitbake also invoke do_task_prepend() before it invokes do_task()? (for example)
Jul 13 15:20:53 <kergoth> _prepend and _append are operations that concatenate a variable
Jul 13 15:21:02 <kergoth> that's true whether the variable is a task or not
Jul 13 15:21:05 <kergoth> see the bitbake user manual
Jul 13 15:22:07 <yates> but this isn't a variable, rather a function
Jul 13 15:22:18 <yates> https://paste.fedoraproject.org/paste/X~mFCEgRGn71HBPvaYRA8A
Jul 13 15:23:31 <kergoth> a function is a variable that happens to have a flag set telling bitbake it's executable
Jul 13 15:23:38 <kergoth> it's all just key=value
Jul 13 15:23:48 <kergoth> same for a task, it's a function with a couple flags set
Jul 13 15:23:57 <kergoth> which is a variable with the func flag set
Jul 13 15:24:13 <kergoth> from bitbake's perspective, a variable is a variable, whether a function or not
Jul 13 15:26:26 <yates> so are you saying do_task_prepend(){foo}, when do_task(){bar} prepends do_task() so that it is now {foo; bar} ?
Jul 13 15:26:53 <kergoth> with a newline instead of a semicolon there, but yes
Jul 13 15:26:58 <kergoth> and bitbake -e will show you exactly that
Jul 13 15:27:21 <yates> ding. i didn't know all this...
Jul 13 15:27:23 <kergoth> again, see the bitbake user manual, it covers the file format
Jul 13 15:30:27 <yates> i have referred to the bitbake manual many times now. i don;'t remember, or have not seen, and do not see now, it stating this information about functions versus variables.
Jul 13 15:30:38 <yates> perhaps i'm missing it.
Jul 13 15:31:27 <yates> i see it now. section 3.4.5
Jul 13 15:33:16 <yates> if there are two prepends on a function, one in a .bb file and the other in a .bbappend file, are they concatenated or does the .bbappend prepend override the .bb prepend?
Jul 13 15:33:48 <kergoth> it's cumulative
Jul 13 15:33:57 <kergoth> _prepend is an operator, like +=, not part of the variable name
Jul 13 15:37:36 * RP can't believe how painful it is to get timing numbers out of unittest :(
Jul 13 15:53:01 <yates> i strongly suggest you modify section 3.1.8 of the bitbake manual to cover the fact that this syntax applies to functions as well as variables.
Jul 13 16:23:42 <RP> The frustrating thing with this parallelisation code is that it assigns tests to queues in advance and assumes test runtimes average out :(
Jul 13 16:33:33 <rburton> oh, nuts
Jul 13 16:33:38 <rburton> that's just dumb
Jul 13 16:34:08 <rburton> have they rejected doing it with a proper queue?
Jul 13 16:34:50 <RP> rburton: the trouble is testtools is retrofitting around upstream unittest so you're stuck with its interfaces
Jul 13 16:35:08 <RP> It may be possible to do this dynamically, not sure
Jul 13 16:35:59 <RP> its the kind of python I'm less good at :/
Jul 13 16:38:41 <RP> rburton: just confirmed a load of the test threads completed and are idle
Jul 13 16:38:56 <rburton> *sigh*
Jul 13 16:39:20 <RP> down to 2 of the 15 threads I think
Jul 13 16:40:09 <rburton> are you using ConcurrentTestSuite?
Jul 13 16:40:20 <RP> rburton: yes
Jul 13 16:40:32 <rburton> would ConcurrentStreamTestSuite be any better?
Jul 13 16:40:50 <RP> rburton: it could solve some problems but not this one
Jul 13 16:41:04 <rburton> shame
Jul 13 16:41:08 <rburton> wasn't sure from the summary
Jul 13 16:41:12 <RP> rburton: with the downside of requiring rewriting everything to use streams
Jul 13 16:41:23 <rburton> ah nuts
Jul 13 16:41:24 <RP> streams aren't in upstream unittest
Jul 13 16:41:30 <rburton> ah ok, fair enough
Jul 13 16:41:32 <RP> afaict
Jul 13 16:58:21 <RP> rburton: We could split up the devtool tests into more modules. I'm going to have a look at that
Jul 13 17:22:54 <khem> rburton: I resent the mesa wayland/egl patch with typo fix in it
Jul 13 19:28:21 <armpit> khem, mesa typo fix ; ) did you see "bild" one
Jul 13 19:29:11 <armpit> in the subject
Jul 13 20:41:36 <yates> so meta-<layer>/swupdate gets applied even if meta-swupdate is removed from bblayers.conf?!?
Jul 13 20:42:18 <yates> (the build's bblayers.conf, that is)
Jul 13 20:52:31 <Crofton|work> shouldn't
Jul 13 21:06:42 <nathani_> how do I force the kernel to rebuild, outside of a cleanall?
Jul 13 21:07:22 <JPEW> nathani_ bitbake -C compile virtual/kernel ?
Jul 13 21:10:23 <nathani_> yeah I think that will do it, thanks!
Jul 13 22:26:31 <yocti> New news from stackoverflow: Building pandoc into yocto core-image-minimal <https://stackoverflow.com/questions/51333347/building-pandoc-into-yocto-core-image-minimal>
Jul 13 23:33:01 <ant_home> khem, fyi https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86520;
Jul 13 23:33:02 <yocti> Bug 86520: was not found.
Jul 13 23:33:19 <ant_home> but it sppears fixed with our gcc8.1.0
Jul 13 23:59:21 <ds2> is there a easy way to find out what is pulling in a recipe?
Jul 14 00:01:22 <kergoth> oe-depends-dot after using bitbake -g
Jul 14 00:01:40 <khem> ds2: yes bitbake -g <image> then it will dump dotty files where you can check for right hand side e.g. search for -> "foo
Jul 14 00:04:53 <ds2> thanks. will try that
**** ENDING LOGGING AT Sat Jul 14 03:00:00 2018