**** BEGIN LOGGING AT Fri Sep 01 03:00:01 2017 Sep 01 03:00:08 hi yahiafarghaly - yes I'm here Sep 01 03:00:58 so if you can't run apt-key then there's a native dependency missing somewhere, presumably it isn't part of our standard apt build Sep 01 03:01:03 I don Sep 01 03:01:19 I don't know if it's a separate package or the result of something being disabled in our apt build Sep 01 03:02:19 apt-key needs gnupg package . so you will notice i have added it to the image recipe but this for the target machine not when the yocto do building stuff Sep 01 03:03:15 native dependancy ? you mean on my machine which is not related to yocto ? Sep 01 06:03:15 If I use the update-alternatives class, and I set the ALTERNATIVE_PRIORITY of one package higher than another, shouldn't I be able to assume that the higher prio one is selected as the default on first boot? Sep 01 06:06:40 Now I am seeing, on first boot, that the alternatives file is correct (the prios are what I set them to) - but the symlink in the rootfs is to the lower prio alternative. Sep 01 06:49:53 i have defined PACKAGE_CLASSES ?= "package_rpm", can i compile any package of meta-debian layer ? Sep 01 07:17:42 I have updated my "oe-core" layer on Pyro from "81246ed classes/staging: change fixme debug message from note to debug" (Jul 20) to "cf1099b connman: Fix for CVE-2017-12865" (Aug 30) and the build breaks. I get: "ERROR: When reparsing /my-image.bb.do_image_cpio, the basehash value changed from to . The metadata is not deterministic and this needs to be fixed." Sep 01 07:18:21 Rebuilding fixes the issue, but the first clean build always fails. Any ideas on how to start debugging this? Sep 01 07:26:45 Hi all, someone knows the best way to give parameters to a recipe ? i tried to add a file in topdir and parse the file during do-install function but the recipe is not rebuild even if the file has been changed. I think do_install[nostamp]"1" should work but it's a litle dirty isn't it ? Sep 01 07:33:50 i have defined PACKAGE_CLASSES ?= "package_rpm", can i compile any package of meta-debian layer ? Sep 01 10:30:02 Are you sure yocto wants to host every single kernel release tar ball on it own at http://downloads.yoctoproject.org/mirror/sources/ ? downloading one gigabyte blobs isn't really useful is it... Sep 01 10:30:19 for example http://downloads.yoctoproject.org/mirror/sources/git2_git.yoctoproject.org.linux-yocto-4.12.git.tar.gz Sep 01 10:30:56 I would host a single git tree for all kernels there, and preferably host it as a git tree instead of a tar ball.. Sep 01 10:40:09 mcfrisk: http://git.yoctoproject.org <-- tell zedd, he has different repos for each major kernel release Sep 01 10:40:28 and that tarball is a tarball of a git repo, as it is the mirror Sep 01 10:41:40 yes, I know. I was just suprised after update to my local poky master branch started downloading gigabytes of data from there. I just built the tree few weeks ago and local caches must have been more uptodate.. Sep 01 10:42:00 due to a new major version, probably Sep 01 10:42:30 The KERNEL_DEVICETREE in imx7dsabresd.conf has a lot of entries. Does/can the kernel actually use others than the one passed by u-boot and the ones that one includes? Sep 01 10:56:48 I removed all except imx7d-sdb.dtb, and it still boots. That's something :) Sep 01 11:02:48 rburton: I guess this is the right approach, a wish list bug: https://bugzilla.yoctoproject.org/show_bug.cgi?id=12027 Sep 01 11:02:49 Bug 12027: minor, Undecided, ---, saul.wold, NEW , Merge linux-yocto git trees to a single one to reduce download times and mirror sizes Sep 01 11:22:05 I am running yocto 2.2.2 on an iMX6UL, trying to get a PAN1026 bluetooth module working through UART2. Sep 01 11:23:08 I am running yocto 2.2.2 on an iMX6UL, trying to get a PAN1026 bluetooth module working through UART2.Can't init device hci0: Invalid request code (56)" Sep 01 11:23:38 Damn scrollwheel... sorry about the spamming Sep 01 11:24:50 I have gotten as far as to get it initialised through the hciattach command using the "any" driver, but when I try to "hciconfig hci0 up" I get the error: "Can't init device hci0: Invalid request code (56)" Sep 01 11:27:01 Does anybody have any hint to where to start looking? Sep 01 11:39:46 Could it perhaps be because the bluez5 tools doesn't work with the imx 4.1.33 kernel as supplied by NXP? Sep 01 11:43:57 hi! a little help to fetch data files from perforce to use in my recipe? (always i am trying to get a proyect *.pro) Sep 01 11:51:09 Hi everyone, someone with time to help a bit? Sep 01 11:52:06 zuribost: Depends on your question ;) Sep 01 11:54:07 I am creating a recipe which needs take data files from perforce repositoy and deploy "as are" in my image. Currently I know how to deploy data from my folders and also I know how to take code from perforce to compile and deploy it Sep 01 11:55:15 I also have question: Is there a way to have wic create an extended partition even if there are only 4 partitions? I'm asking because in 2.1 it created the 4th partition as extended and to keep backwards compatibility I would like to keep it that way. Sep 01 12:07:47 The change that introduced that change was: https://github.com/openembedded/openembedded-core/commit/485315dc170e29962a8848db38db73abafd0586e Sep 01 12:10:10 bachp: ed21 is the maintainer but it looks like partition 4 being extended was a bug Sep 01 12:12:23 rburton: Maybe but as it was like this in krogoth we now expect it to be that way and it would break our update process Sep 01 12:13:56 I can understand that not putting partition 4 into the extended partition is the better default but I'm hoping I can somehow configure it to still put it into the extende part without writing our own imager plugin :) Sep 01 12:21:41 Hi. Does anyone know how to execute a task right before reboot/shutdown/halt with systemd (in krogoth)? Sep 01 12:22:25 I am trying to get my first built attempt to work on a Orange-pi-zero. The image doesn't boot although it built without any issues. I looked at the boot.scr file on the SD card (with vim in the debain VM it was built with) and there is a few lines of ASCII jibberish before the normal text that you would expect to be there. Is it supposed to be there or is that not a correct way to view the file? Sep 01 12:37:07 Is there a way to display the dependencies of a particular task in bitbake? Sep 01 12:40:15 BubuIIC: -g will dump a .dot file for you Sep 01 12:45:30 rburton: so if the resulting task-depends.dot is empty I did something wrong specifying the dependencies. Sep 01 12:45:50 That at least confirms my suspicions, thanks Sep 01 14:27:38 RP ping Sep 01 14:28:33 armpit: pong Sep 01 14:29:14 RP, has the latest pyro been run on the AB? Sep 01 14:29:21 armpit: yes Sep 01 14:29:24 armpit: all green Sep 01 14:29:25 k Sep 01 14:29:43 armpit: I tested peter's changes, dropped texinfo, retested and merged Sep 01 14:30:29 armpit: looking forward to testing your morty fix Sep 01 14:31:35 I am planning on firing off a build this weekend unless you queued on up Sep 01 14:32:02 there are a few more patches ( security) I need to merge Sep 01 14:32:03 armpit: fine by me. halstead will be doing maint, then you can use autobuilder.yp.org Sep 01 14:32:10 k Sep 01 14:32:42 I'm running a -next on the other one having spent all day chasing a nasty corruption causing logic error or two :( Sep 01 14:32:53 who do I make requests for patchset system ie allowing to change states Sep 01 14:33:26 armpit: lamego may be able to help? Sep 01 14:33:32 or lsandov Sep 01 14:33:47 I can login and all but would like to use the other features Sep 01 14:33:52 k Sep 01 14:34:15 armpit: or paule but I think the others are more likely to be around shortly Sep 01 14:34:24 reading.. Sep 01 14:37:29 armpit: Are you talking about editing patches/series from the patchwork UI? Sep 01 14:38:24 not editing, but allowing state changes. new to other state Sep 01 14:39:03 i can create bundles Sep 01 14:40:48 armpit: that is a permissions case. By default, a patchwork user can edit only patches submitted by himself. from others if he/she has maintainers permissions for a given patchwork project (for example, oe-core). Sep 01 14:41:29 armpit: Do you want to change status for patches submitted fby other user? Sep 01 14:41:35 *by Sep 01 14:42:36 lamego: armpit handles some of the stable branches so I suspect the answer is yes Sep 01 14:42:59 it would be nice. would help me in my maintenance process for core and meta-oe Sep 01 14:44:21 armpit: RP: then I can give you such permissions for the patchwork projects that you prefer. What would those be? Sep 01 14:44:55 armpit: oh! I just saw your answer Sep 01 14:45:21 armpit: I 'll work on that in a moment and get back to you. Sep 01 14:45:36 k Sep 01 15:06:26 * kergoth yawns Sep 01 15:23:16 kergoth: some of the errors we're seeing with BB_SERVER_TIMEOUT scares me :/ Sep 01 16:51:15 halstead: there is a master-next ready for testing after the current one on yocto.io. Assuming you'd like to do the weekly maint before that one starts though. Sep 01 17:07:09 RP, I can do it after lunch meetings. I say go ahead and queue it up now. Sep 01 17:11:53 halstead: ok, thanks Sep 01 18:16:09 Hello everyone! Sep 01 18:16:12 Happy Friday! Sep 01 18:17:19 I have nginx installed, as well as a working config. How do I set the nginx service to start automatically? I tried systemctl enable nginx to my recipe, but it failed. Sep 01 18:19:28 tcpdump: did you see http://www.yoctoproject.org/docs/current/mega-manual/mega-manual.html#var-SYSTEMD_AUTO_ENABLE? Sep 01 18:19:29 https://stackoverflow.com/questions/32902003/enable-systemd-service-in-yocto-build Sep 01 18:19:52 OK - cool!! I found the same thing but elsewhere. Sep 01 18:21:01 So would it be SYSTEMD_AUTO_ENABLE nginx= "enable" ? Sep 01 18:21:51 SYSTEMD_AUTO_ENABLE_nginx = "enable", if anything. Depends on how your recipe looks like, though. Sep 01 18:23:06 if you use the nginx recipes from meta-webserver, it should work this way Sep 01 19:11:52 RP you still around? Sep 01 19:53:37 why am i getting a "Parsing recipes..Traceback"... when I devtool modify -x linux-variscite src/linux ? Sep 01 20:03:20 the tool should NEVER crash like this Sep 01 20:04:29 it should not can you provide more information about crash ? Sep 01 20:09:22 khem: does such a crash generate a .log somewhere? Sep 01 21:03:55 yates you might check tmp/logs/cooker Sep 01 21:06:39 khem, hi, giving a try to gcc 7.2 vs armv5te right now Sep 01 21:21:16 ant_home: ok. I am sorry I dont have much data to help you Sep 01 21:23:38 khem: ok, thanks. Sep 01 21:34:26 khem, new binutils as well Sep 01 21:43:40 armpit: back now, back again in a bit Sep 01 21:46:30 RP, there was an issue with uninative in pyro, is there a version I should update to? Sep 01 21:47:20 ah, just say your email Sep 01 21:48:48 armpit: the latest basically Sep 01 21:49:09 k, will do Sep 01 21:52:08 RP have a good weekend Sep 01 22:20:49 * armpit hmm uninative is latest Sep 01 22:24:38 uni native ... hmmmm..... Sep 01 22:24:45 1 native. Sep 01 22:45:48 * armpit launches morty-next build Sep 01 22:50:35 armpit: they may have tested before the last upgrade Sep 01 22:55:03 maybe. I notice you pull in the latest on july 30th Sep 01 23:04:42 armpit: I'm not sure when it got pushed though, may have lain in -next for a while **** ENDING LOGGING AT Sat Sep 02 03:00:01 2017