**** BEGIN LOGGING AT Thu Jan 28 02:59:59 2016 Jan 28 10:25:15 IIRC it's to get the layername/LAYERDIR of the base .bb file, but I cannot find the name. Any can give me a pointer how to find it? Jan 28 13:17:30 Hi. Does somebody know if there is a way to install all the dependencies of a package without building the package itself? Jan 28 13:18:54 I would like to have an image that includes all the packages mentioned in the DEPENDS variable of the my-app.bb recipe but I don't want to build my-app. Jan 28 13:21:54 /join #qt Jan 28 13:21:59 sorry* Jan 28 14:40:47 booo Jan 28 14:55:23 trick or treat ? Jan 28 14:57:36 trick, selftest blew up Jan 28 16:08:49 fsdun: ${@bb.utils.get_file_layer('${FILE}', d)} will get you the layer name for the file, but not the layerdir Jan 28 16:08:58 fsdun: meta-mentor provides it already, though.. Jan 28 16:09:01 fsdun: https://github.com/MentorEmbedded/meta-mentor/commit/a9bb837469ebff01397c3334be3f564e43e0fb08 Jan 28 16:29:22 Is there a way to pass through something to git for the bitbake fetcher? Specifically, I wanted to set the credential.helper to cache Jan 28 16:31:05 Hi everyone, Jan 28 16:31:08 I'm trying to bbappend a recipe already "bbappended" by another layer. Jan 28 16:31:23 My layer's priority is higher and is shown after the other one when I type 'bitbake-layers show-appends'. Jan 28 16:31:29 Unfortunately, it's not taken into account (ie, when I change the file used by my .bbappend, the image is not rebuilt) Jan 28 16:31:43 jedix: set FETCHCMD_git = "git -c core.fsyncobjectfiles=0 -c credential.helper=cache" Jan 28 16:31:54 jedix: the fsync thing is what's in the default value of FETCHCMD_git when it's unset Jan 28 16:32:00 afaik that should work, anyway Jan 28 16:32:18 olivier___: are some of the appends using % and others not? Jan 28 16:32:22 thanks kergoth Jan 28 16:32:28 olivier___: try -c clean on the package, and then try again Jan 28 16:32:32 jedix: np Jan 28 16:32:51 CTtpollard: The other one yes, and so does mine. Jan 28 16:33:07 kergoth: The other one yes, and so does mine. Jan 28 16:33:35 now, of only the parsing recipies message didn't trample the input request :) Jan 28 16:34:24 that doesn't surprise me, you'd really need to use a credential helper that pops up in X or something else out of band, probably Jan 28 16:34:50 yeah Jan 28 16:35:21 CTtpollard: I have another recipe with the "same structure". If I modify a file used, 'bitbaking my image' automically rebuild the package Jan 28 16:36:28 FYI, the recipe is ".../recipes-core/init-ifupdown" Jan 28 16:38:04 olivier___: the only issue i know of with appends not obeying priority order was a bug relating to non-deterministic ordering between % and non-% appends, which was fixed in poky commit 3cb87724a5b21550e99c18e97b665d6604bb2fa9 (which was bitbake commit f980f060cd0d1e7fe5011f3c325c1b254f05eccf) Jan 28 16:38:25 beyond that, i don't know of anything that would cause the behavior you described, unless your append isn't being parsed at all Jan 28 16:38:39 did you check bitgbake -e recipename to make sure it was parsed? it will show what files were parsed in the top of bitgbake -e output Jan 28 16:39:54 quick question on % wildcard appends, if I have a %.append and the layer with the highest priority has two recipes for the append, which will the append apply to? Jan 28 16:40:11 i.e, two different version number recipes Jan 28 16:40:22 % is a wildcard Jan 28 16:40:26 it will apply to any version Jan 28 16:40:47 so whichever version you end up preferring / building, it will apply Jan 28 16:41:21 if there's two versions in the layer, how does it determine which? Jan 28 16:42:22 kergoth: So, the file is correctly listed. Jan 28 16:43:06 it will also show the history of changes to the variable in question, so you can see if it was set by this and then overridden by that Jan 28 16:46:45 kergoth: Ok, I'll check. Thanks ! Jan 28 17:15:51 moin Jan 28 17:37:55 anyone using imx6 boards here? or another board with a viv gc880 on it Jan 28 17:38:20 are you seeing really bad gpu performance? Jan 28 17:38:41 * nerdboy hasn't poked at the gpu much yet Jan 28 17:39:24 fbdev on earlier 4.x worked okay Jan 28 17:39:57 which board? Jan 28 17:40:04 not messing with vendor-y blobs/forks much really... Jan 28 17:40:22 wandboard/udoo/cubox Jan 28 17:40:36 we are seeing like 20-30fps on simple examples, its ridiculous Jan 28 17:41:14 I wanted to try newer viv drivers but freescale don't allow listing the download directory :< Jan 28 17:41:20 with what combination of kernel/rootfs? Jan 28 17:41:43 and what hardware? Jan 28 17:41:44 http://www.nxp.com/lgfiles/NMG/MAD/YOCTO/imx-gpu-viv-5.0.12.p7.1-hfp.bin Jan 28 17:41:47 erm Jan 28 17:42:24 Linux imx6dlsabreauto 3.10.17-1.0.2_ga+g33597e3 #1 SMP PREEMPT Fri Jan 22 12:03:50 PST 2016 armv7l GNU/Linux Jan 28 17:44:13 are you building fsl-community-bsp stuff? Jan 28 17:44:21 what branch? Jan 28 17:46:29 3.10.17 looks old, even for a vendor fork... Jan 28 17:46:48 I don't think we are Jan 28 17:51:32 what are you using? Jan 28 17:53:23 meta-fsl-arm? Jan 28 17:53:38 i got 3.14.28-r0-udooneo on jethro the other day Jan 28 17:53:51 but that's a fork of a fork even Jan 28 17:55:23 whats your version of imx-gpu-viv- Jan 28 17:55:57 the old one is 3.14.28-ish for most devices Jan 28 17:56:38 you could try linux-fslc-4.1 Jan 28 17:57:23 imx-gpu-viv_5.0.11.p7.1 in jethro Jan 28 17:57:54 i would try jethro or master Jan 28 18:01:45 have you looked at robert's sabreqsd page? Jan 28 18:01:48 https://eewiki.net/display/linuxonarm/i.MX6q+SABRE+Board+for+Smart+Devices Jan 28 18:41:00 So, how do the SDK targets decide which packages to build for the MACHINE sysroot, and which packages to build for the development sysroot? Jan 28 18:43:12 chicken bones Jan 28 18:44:11 wyrm: what exactly do you mean by 'development sysroot'? Jan 28 19:39:21 Hi guys. Question regarding SRCREV = "${AUTOREV}" Jan 28 19:40:15 the 1.6 manual seems to advocate a fixed SRCREV in the recipe and applying floating SRCREVs in local.conf. Is that still true in more modern poky? Jan 28 19:40:57 i.e. is it more conventional for a published layer to use AUTOREV or not? Jan 28 19:44:43 kergoth: The sysroot that is compiled for the ... development host? The workstation on which stuff is built, to be later installed and run on the destination platform. Jan 28 19:45:08 TOOLCHAIN_HOST_TASK defines host stuff (nativesdk recipes), TOOLCHAIN_TARGET_TASK defines target stuff Jan 28 19:45:12 those correspond to the two sdk sysroots Jan 28 19:45:25 If I'm reading the documentation right, Yocto calls that the TARGET platform. Jan 28 20:11:22 kergoth: too bad wget can't use anything like git's cache Jan 28 20:11:41 I might have to write something to handle wget, in which case the git part will be trivial Jan 28 20:59:31 bengardiner: published layers using autorev are prone to break Jan 28 20:59:49 recomendation is to use static revs in recipes Jan 28 21:00:02 agreed -- I'm glad to hear it Jan 28 21:00:11 if you want a "bleeding edge broken" mode, have a separate file that applies autorev Jan 28 21:00:13 rburton: thanks Jan 28 21:00:29 poky used to have a poky-bleeding which did this, it's less useful these days as it doesn't build much at all Jan 28 21:01:03 rburton: cool -- it's like you're in my head :) I was going to remove all the AUTOREV in the recipes and add an .inc like bleeding :) Jan 28 21:04:30 * nerdboy passes out foil hats Jan 28 21:05:03 bengardiner: beware of Crofton|work ... Jan 28 21:06:28 nerdboy: what's that? "Crofton|work" ? Jan 28 21:07:09 put on your foil hat you'll be fine... Jan 28 21:08:19 * bengardiner puts on foil hat Jan 28 21:09:32 autorev is "interesting" in a real project env... Jan 28 21:10:06 someone always forgets to push their crap Jan 28 21:37:58 nerdboy: or, pushes crap Jan 28 21:40:51 rburton: I'm not sure if it's going to work or not cause they have strange SRCREV_XXX where XXX is not PN ; but I'm gonna give it a shot https://github.com/IrdetoServices/meta-rdk-belvedere/commit/4a5c222bd252488906f36d786b3d335f7971ba90 Jan 28 21:42:10 hm Jan 28 21:42:23 well for recipes where there's just one entry in SRC_URI, you can remove all the name stuff Jan 28 21:43:00 thanks, now I even have a backup for when my testing blows up :) Jan 28 21:43:10 rburton: it's more exciting the other way Jan 28 21:43:18 ! :) Jan 28 21:44:01 haha Jan 28 21:44:17 more hair-pulling trying to figure out what regressed/didn't get fixed anyway... Jan 28 21:44:57 blimey andre is on a roll this week Jan 28 21:46:31 blimey? Jan 28 21:47:07 * nerdboy definitely needs to visit rburton's corner pub Jan 28 21:49:04 warning: britishisms Jan 28 22:01:52 pretty sure that's the first time i've seen it used in a sentence... Jan 28 22:03:07 * nerdboy should use "groovy" in sentence more often Jan 28 22:03:21 nobody says groovy Jan 28 22:03:50 that would be the point... Jan 28 22:04:46 although groovin' made a minor comeback in college Jan 28 22:43:19 is there an easy way to run rm_work on all built recipes (or all recipes the image depends on), to make some space after-the-fact (I'm constantly switch rm_work on and off to debug things while avoiding filling my disk even more...) Jan 28 22:49:47 can gcc libssp be disabled from compiling? Jan 28 23:37:18 where can i get my poodle clipped in burbank? Jan 28 23:53:46 * nerdboy fears everyone is off listening to frank now... **** ENDING LOGGING AT Fri Jan 29 02:59:59 2016