**** BEGIN LOGGING AT Thu Jan 21 02:59:59 2016 Jan 21 09:10:50 good morning Jan 21 09:59:12 hi mckoan Jan 21 10:07:53 hi bluelightning Jan 21 16:05:05 Hello Jan 21 16:05:19 There is a missing dependency in meta-openembedded Jan 21 16:05:40 I am not sure the best way to add it Jan 21 16:06:39 ntp misses the rdepends "perl-module-getopt-std" Jan 21 16:07:32 RDEPENDS_${PN} += "perl-module-getopt-std" Jan 21 16:07:36 in the ntp recipe Jan 21 16:09:18 ok, because the ntp bb starts with a ntp.inc Jan 21 16:18:44 hmm maybe the RDEPENDS_${PN}-utils += "perl" does the job Jan 21 16:24:54 kergoth: any thoughts on the data_smart patch I sent to the bitbake list? Jan 21 16:29:19 RP: oh, hey, that'll fix the long-standing behavior where use of a ${} in inline python would explode so we had to switch to d.getVar()s? Jan 21 16:29:38 looks reasonable to me. not pretty, but we're stuck for now Jan 21 16:29:39 hmm Jan 21 16:30:04 s/use of a \${}/use of a \${} for a undefined var/ Jan 21 16:31:16 I'd still like to see use of undefined ${} raise an error, as it'd catch typos easily and it's not hard to add default values, but then add the ability to escape it for shell passthrough, ie. \\${}.. but then we'd have to take that into account when we want \ to pass through.. Jan 21 16:31:22 afk Jan 21 16:32:37 kergoth: in image.bbclass, I'm 'abusing' this to defer expansion of certain variables too Jan 21 16:33:29 kergoth: I see the image.bbclass code as a stepping stone to something more sane for image tasks so hopefully it won't be around for too long Jan 21 16:35:32 kergoth: ultimately I agree about the undefined vars. The question is whether that patch makes things better or worse for now... Jan 21 16:38:58 I think it's worth applying as is, it's an improvement to a fairly common pain point Jan 21 16:40:40 kergoth: agreed, that was my thinking too. I prefer it to the patch proposed to 'fix' image.bbclass for example Jan 21 16:40:51 agreed. Jan 21 16:44:15 kergoth: as a bonus, I think I can clear up some of the hacks in bitbake.conf with this... Jan 21 16:44:30 kergoth: http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wip&id=aea168f014cb3cd164bea4a7d9560f9af10a802a Jan 21 16:44:37 it'd be nice to see the PKGE nastiness go away Jan 21 16:44:47 right, that :) Jan 21 16:44:59 kergoth: that patch has one bit of reversed logic but you get the idea Jan 21 16:45:07 yeah Jan 21 17:47:34 OK, in an effort to avoid the FOSDEM sickness, I had a flu shot Jan 21 17:48:53 Crofton|work: sounds like a good idea Jan 21 17:49:12 just found out it takes two weeks for full effect, so a bit late :0 Jan 21 17:49:22 Crofton|work: I watched someone wipe their nose and then immediately shake someone's hand last year, no wonder people get sick Jan 21 17:50:12 yeah, with the weather, going in and out etc hard to cope Jan 21 18:25:12 I'm running afoul of the automatic removal of shadow when you don't have package_management enabled, is there a knob somewhere that I'm missing? Jan 21 18:25:36 It doesn't look like it from looking at rootfs.py... Jan 21 18:32:25 smurray: we did recently change it so it only removes them now if your rootfs is read-only, rather than the fact that you don't have a package manager Jan 21 18:33:09 (recently as in November in the master branch) Jan 21 18:33:45 bluelightning: thanks, I just thought to check master instead of jethro and saw that Jan 21 18:34:53 bluelightning: slight issue is I'm currently upgrading from dizzy to fido, may just backport that change locally for now Jan 21 21:08:41 amyone remember what the IMAGE_FEATURE qt4-pkgs did? Jan 21 21:08:42 ERROR: gnuradio-demo-image was skipped: 'qt4-pkgs' in IMAGE_FEATURES is not a valid image feature. Jan 21 21:09:51 -FEATURE_PACKAGES_qt4-pkgs = "packagegroup-core-qt-demoapps" Jan 21 21:09:58 installed the qt4 demos Jan 21 21:10:44 almost certainly that was more than you needed Jan 21 21:12:42 it was called -demo-image :) Jan 21 21:12:50 but I suspect I can drop them with no one noticing Jan 21 23:13:44 at4-pkgs from image. **** ENDING LOGGING AT Fri Jan 22 02:59:58 2016