**** BEGIN LOGGING AT Fri Feb 19 02:59:58 2016 Feb 19 08:28:52 good morning Feb 19 13:23:33 <_gpg_> hello every body Feb 19 13:24:23 <_gpg_> any one can explain me the following error please ? QA Issue: libQt5Widgets.la failed sanity test (workdir) Feb 19 13:29:02 the rest of the message explains the error Feb 19 13:29:11 but martin sent a fix to oe-devel recently Feb 19 13:39:00 <_gpg_> rburton1: thanks i was reading insane.bbclass and i'm quite confused, is there any change to see martin's fixes ? Feb 19 13:39:17 <_gpg_> rburton1: i realy need to test my application :( Feb 19 13:39:35 you'll need to update meta-qt5, or apply the patch if its not already merged Feb 19 13:39:42 i don't use qt5 so can't help more really Feb 19 13:40:28 <_gpg_> rburton1: i'd prefer to apply the patch, just wondering where to get it from Feb 19 13:40:46 oe-devel mailing list Feb 19 13:45:34 _gpg_: you can use jansa/master branch which already has all fixes I've sent yesterday Feb 19 13:46:04 <_gpg_> JaMa: i'm using Jansa/5.6 (since i need a specific stuff from Qt 5.6) Feb 19 13:47:15 <_gpg_> JaMa: and i'm dumb cause it looks that master already uses QT 5.6rc Feb 19 13:48:16 <_gpg_> JaMa: THANKS SO MUCH Feb 19 13:48:25 <_gpg_> JaMa: thank you so much Feb 19 13:49:27 jansa/master (and -5.6) only has a bit newer revisions from 5.6 branches Feb 19 13:51:50 Hi all. Is there any reason the contrib subdirectory of scripts directory isn't exported to PATH in oe-init-buildenv? Feb 19 13:52:07 Are those scripts considered unstabled/ -tested etc.? Feb 19 14:00:32 pretty much Feb 19 14:00:37 historical, really Feb 19 14:00:50 if there's one that you think should be moved, send a patch Feb 19 14:16:50 rburton1: Well i kind of like the ddimage script Feb 19 14:17:12 Although it doesn't do a hell of a lot it's just conveneint Feb 19 14:17:20 convenient* Feb 19 14:18:36 * rburton1 prefers image-usb-stick Feb 19 14:38:06 why every build executes NOTE: recipe signing-keys-1.0-r0: task do_export_public_keys: Started Feb 19 14:38:12 and rm_work after that Feb 19 14:38:19 even when I'm not using signing-keys? Feb 19 14:40:39 because you built os-release Feb 19 14:40:41 yes, i know Feb 19 14:41:33 there's a patch from randy on the list that should fix that, but i've not looked at it yet mysef Feb 19 14:41:40 i see markus gave it a review already Feb 19 14:43:15 rburton1: i think it'd be good if you took a quick look, too :) Feb 19 14:43:48 yeah Feb 19 14:43:54 inbox is a bit top-heavy right now Feb 19 14:45:28 rburton1: ty will cherry-pick it to my fork Feb 19 14:46:41 rburton1: i can guess that :/ Feb 19 16:02:31 Crofton: thanks for the SUMMARY clean up patch(es) Feb 19 16:02:43 rburton1: thanks for the python.elementree patch Feb 19 16:57:32 Hi all. Yocto newbie here. Building x86_64 sato live-boot image from master with linux-yocto-dev 4.4 kernel, X server fails to start as tmpfs mounts are read-only. No aufs in kernel. Is this expected/known-issue? Feb 19 17:27:31 hi, is there an easy way to get optparse for python with Yocto? I can only see argparse. Feb 19 17:34:25 when I write "bitbake python", I cannot seem to find a "big" python package containing this, but I could be doing something silly. Feb 19 17:36:21 that's.. weird Feb 19 17:36:39 poky/meta/recipes-devtools/python/python-2.7-manifest.inc Feb 19 17:36:39 254:FILES_${PN}-textutils="${libdir}/python2.7/lib-dynload/_csv.so ${libdir}/python2.7/csv.* ${libdir}/python2.7/optparse.* ${libdir}/python2.7/textwrap.* " Feb 19 17:36:43 from a quick grep Feb 19 17:36:48 why the hell would it be put in textutils? Feb 19 17:36:53 that's a weird package to put it in Feb 19 17:37:11 lpapp: oe-pkgdata-util is your friend, as is grep :) Feb 19 17:39:17 ah, right, thank you very much! Feb 19 17:40:19 np Feb 19 17:40:40 * kergoth still doesn't get the textutils thing, but at least it's somewhere :) Feb 19 17:40:49 true. Feb 19 17:41:21 there's also a python-with-batteries-included package that depends on a bunch of other stuff Feb 19 17:41:30 I don't exactly recall its name at the moment, though. Feb 19 17:42:32 iirc its python-modules or somesuch. really, 'python' should pull in all the modules that come with python itself.. Feb 19 17:51:11 RP: any thoughts on noor's email about dev packages pulling in additional non-dev packages due to the depchains using DEPENDS rather than RDEPENDS? Feb 19 17:51:41 (re: "[OE-core] some dev pkg bring extra target packages when installed in rootfs") Feb 19 17:54:48 kergoth: hmm. there are two ways to solve this. One is to do what you describe, the other is to add RPROVIDES to the -dbg packages Feb 19 17:55:15 I guess option three is a dbg package per package containing binaries Feb 19 17:55:31 kergoth: rburton was looking at this too FWIW Feb 19 17:55:40 in this case it's -dev causing issues rather than dbg, but both are potentially of concern Feb 19 17:58:32 kergoth: right, its the same code iirc Feb 19 17:58:46 * kergoth nods Feb 19 18:00:19 i think we're stuck when it comes to dev depchains, it's all or nothing, since -dev can't be per-binary-package, by definition it's recipe-wide Feb 19 18:00:21 * Noor finds his email getting attension Feb 19 18:00:22 but dbg we could improve Feb 19 18:16:03 kergoth, RP: i was thinking adding provides would be easier Feb 19 18:18:25 The issue we were hitting was basically adding dev-pkgs didn't *only* add dev packages, which is potentially confusing to the user, if something wants libreadline which wants libcurses, and installs readline-dev, that sucks in all of ncurses as well, not just the main lib, since ncurses-dev depends on ncurses. but i don't think there's any way to avoid this, nature of the beast Feb 19 18:22:36 isn't installing ${PN}-dev package the statement of "I want a development environment that allows me to develop against ${PN}", which includes libraries ${PN} depends on? Feb 19 18:23:04 So that sounds like a non-issue to me; or am I misunderstanding things? Feb 19 18:23:28 well, it's not just libraries, of course. it's everything. extra binaries, which suck in other deps, etc Feb 19 18:23:40 but yeah, probably not much to be done about it Feb 19 18:25:09 Yes, but if I want to build against X, and that executes a binary provided by Y at runtime, and after linking and running my code I notice that I have to install Y manually to make my development environment complete, *then* I'm going to be confused. Feb 19 19:07:29 bitbake warns if SDK_PACKAGING_FUNC is ":". if its "", we get an unpleasant ShellSyntaxError traceback Feb 19 19:07:40 i'm thinking we need an official way to bypass create_shar without getting warnings or errors Feb 19 22:11:10 moto-timo, no problem. **** ENDING LOGGING AT Sat Feb 20 02:59:58 2016