**** BEGIN LOGGING AT Fri Aug 18 03:00:02 2017 Aug 18 10:15:50 Hello to the community. Aug 18 10:16:17 Question: where I can find python-rpm package? Aug 18 10:16:31 Asks me for adding smart packager Aug 18 10:17:15 I went to find it in meta/recipes-devtools/phyton, but it is not there??? Aug 18 10:18:24 so I have a _svn.bb and a _1.0.bb. For this build I'd like bitbake to use the _1.0.bb. How do I tell it to use that recipe? I've tried to set the preferred version to -1 without success Aug 18 10:20:30 So, in python-smartpm_git.bb I have 4 dependencies: DEPENDS = "python rpm gettext-native python-rpm" Aug 18 10:20:55 I have first three, but NOWHERE to find python-rpm??? Aug 18 10:23:02 nobody: Hi, I was doing something like this last week I think. Aug 18 10:23:56 nobody: let me check my notes and I will answer unless someone else doe sfuirst. Aug 18 10:30:31 nobody: Sorry I was mistaken. I was getting a a Python library added to a source image build. I found that there were missing dependencies in the Python library's .bb file so it did not cause the system to include all the required Python Libraries in the image and packge lists. That is not what you are trying to do, is it? Aug 18 10:31:59 T_UNIX: _svn.bb and _1.0.bb are not two versions of the same recipe. Are they? Well assuming they are for now, are they provided by the same layer or two different layers? Aug 18 10:32:46 T_UNIX: what output do you get for `bitbake-layers show-recipies Aug 18 10:34:01 T_UNIX: what output do you get for `bitbake -s | grep ` ? This should list the version being used first then other non-used high versions. It hides lower versions. The other command I sent should tell you if the system can see the recipies, even if it is not using them. Aug 18 10:35:24 TafThorne: they are part of the same leyer Aug 18 10:35:36 i.e. one is a dev version, while the other one is a release Aug 18 10:35:56 T_UNIX: earlier this week I had a _2.9.1.bb and a _3.1.bb version of a recipie. Only the 2.9.1 version was being used. The two recipies were in different layers. Layer priority is more important that recipie version number. The layer that provided _3.1.bb was a lower layer priorty than the layer providing _2.9.1.bb so the newer version was never used. Aug 18 10:36:17 I use a .inc to sort out common stuff. Aug 18 10:36:45 and use the Filename + ${PV} within the recipe for the releases Aug 18 10:38:29 bitbake-user-manual might be missing the bit we need here "We need to document AUTOREV and SRCREV_FORMAT here." Aug 18 10:42:00 T_UNIX: maybe you should be using BBVERSIONS inside a single recipie file? Aug 18 10:42:08 TafThorne: `bitbake -s | grep ` lists the build order. I thought that increasing or decreasing the PREFERRED_VERSION was relative to that sorted list. Aug 18 10:44:13 TafThorne: thanks. That might be the way to go :) Aug 18 10:44:49 T_UNIX: I thought PREFERRED_VERSION_ would do that too. It is limited by the layer priorty. It may be limited by other things too. I am quite new to all this. People seem more active on here in the afternoons so perhapse check back then for more pointers. Aug 18 10:45:17 T_UNIX: You are welcome. Good luck sorting out your orders of precedence. Aug 18 11:11:52 nobody: Sorry I was mistaken. I was getting a a Python library added to a source image build. I found that there were missing dependencies in the Python library's .bb file so it did not cause the system to include all the required Python Libraries in the image and packge lists. That is not what you are trying to do, is it? Aug 18 11:12:32 Thank you for the answer, Thorne. Aug 18 11:13:00 I am not sure, if this package: phyton-rpm is not obsolete??? Aug 18 11:13:53 I'll try to again to activate my smart manager, and see if I can make it work without that. Aug 18 11:16:04 Aug 18 11:19:31 nobody: phyton-rpm or python-rpm or python and then just set the system to getnerate prm? Aug 18 11:22:27 On the target QEMU x86_64 I gave the command: smart channel --add all type=rpm-md baseurl=http://127.0.0.1:8000/all Aug 18 11:23:35 Then target (QEMU) asks me: include this channel? after Y, it prints: error: 'rpm' python module is not available Aug 18 11:25:37 So I have all three packages I need: rpm, python, but NOT python-rpm, I commented this one out! Aug 18 11:27:08 So, in meta/recipes-devtools/phyton/python-smartpm_git.bb I have 4 dependencies: DEPENDS = "python rpm gettext-native python-rpm" Aug 18 11:27:36 last I could not track, neither in poky source, neither on the net??? Aug 18 11:34:15 And, TafThorne, smartpm was also not inculed by default in poky, although it says that it was included. Aug 18 11:34:37 I was able to track it, and add locally. Aug 18 11:40:30 http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/python Aug 18 11:42:40 No signs of smartpm and python-rpm in this directory?! Aug 18 11:45:28 http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/python?h=morty Aug 18 11:45:37 Here we see smartpm... Aug 18 11:45:48 In morty Aug 18 11:47:17 And here is .smart.bb content in morty: http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/python/python-smartpm_git.bb?h=morty Aug 18 11:53:03 And here is... Morty has it, python-rpm??? http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-devtools/rpm/?h=morty Aug 18 12:01:37 As I am thinking more and more, TafThorne, something here, in master branch, is NOT done correctly, so smartpm (as is) can NOT be used with poky 2.2 and poky 2.3 Aug 18 13:26:18 nobody: smartpm was replaced with dnf in 2.3 Aug 18 13:48:34 RP: ping. 4.12 will be out in a series today. It has looked good for a day now, just doing extra testing with muslc, etc. Aug 18 13:49:27 RP: but my other feature is going to miss, but I'll send a RFC. I can't figure out what has changed I'm now getting hit by: Aug 18 13:49:30 | /home/bruce/poky/build/tmp/work/qemuarm64-poky-linux/linux-yocto/4.12.7+gitAUTOINC+6377a5aee6_918fc2fbdf-r0/temp/run.do_kernel_configme.8161: 122: /home/bruce/poky/build/tmp/work/qemuarm64-poky-linux/linux-yocto/4.12.7+gitAUTOINC+6377a5aee6_918fc2fbdf-r0/temp/run.do_kernel_configme.8161: Bad substitution Aug 18 14:02:53 zeddii: what is on line 8161? Aug 18 14:04:04 zeddii: sorry, that is the pid, line 122 Aug 18 14:04:27 an inline python call. I think I just managed to figure out a bit more. Aug 18 14:04:47 I pushed core fragment support to kernel.bbclass and I'm just trying to set a variable in my kernel-yocto stuff to use it. Aug 18 14:05:21 the variable in question is a list of fragments. and yes, I think it is expanding into something python doesn't like. I'm evil and mixing shell and python, when really, I need to just convert the routine to python. Aug 18 14:06:18 zeddii: sounds like you have a handle on what is going on then :) Aug 18 14:06:24 yah. I'm good. Aug 18 14:06:34 as in, I'm not thinking that inline python calls are busted. :D Aug 18 14:06:40 that was never really an option. Aug 18 14:06:46 but when you are drowning .. ;) Aug 18 14:06:55 I'll go back to checking on 4.12 and cleaning that up. Aug 18 14:19:18 zeddii: I know all about drowning. Keeping patches moving is about all I can do atm Aug 18 14:39:16 * zeddii has learned he cannot call inline anonymous python in a loop. Aug 18 15:03:54 zeddii: heh, inline python is run when it's expanded, not when the task is run :) Aug 18 15:08:11 Hello Aug 18 15:22:16 kergoth. heh. yep. learning by failing :D Aug 18 15:37:26 RP: I can't find an answer in the bitbake manual. can taskA set a variable that taskB can see ? i.e. via d.set/getVar ? Aug 18 15:37:53 zeddii: nope, each task gets its own datastore Aug 18 15:38:04 aha Aug 18 15:38:18 so I'll go via a file or another method. Aug 18 15:38:27 specifically *tasks*, not functions. i.e. a prefunc can set soemthing the main task can see, and multple functions called by the main task can interact that way, but separate tasks, no Aug 18 15:38:47 cool. makes sense, too many ways to abuse it if they could. Aug 18 15:38:55 kinda like I'm trying to do ;) Aug 18 15:39:36 zeddii: exactly ;-) Aug 18 15:40:08 so I'll just channel the fragments out via something in workdir and pick it up via that in the core bbclass. Aug 18 15:40:16 learned something new today :P Aug 18 17:05:59 Any one suggest me, following testsuite to test selinux ? Aug 18 17:06:00 https://github.com/SELinuxProject/selinux-testsuite Aug 18 17:06:21 in yocto ? Aug 18 17:28:15 Puru: look at ptest for selinux package Aug 18 18:27:05 just when I thought I was done. qemuarm64 blows up on 4.12 and ssl enanglements Aug 18 18:43:04 welcome to the muddy waters Aug 18 19:17:04 RP .io is ready for a build. Still working on debian8 though. Aug 18 22:27:40 Hey all. I'm trying to get some generic files copied onto a system image. I followed an example I found archived in the mailing list, but it's throwing an error I don't understand. Aug 18 22:27:50 https://lists.yoctoproject.org/pipermail/yocto/2015-December/027681.html <--example Aug 18 22:27:56 ERROR: core-image-firebird-1.0-r0 do_rootfs: scripts2 not found in the base feeds Aug 18 22:28:56 https://pastebin.com/NYjr6Rh0 <--scripts2.bb Aug 18 22:29:19 Any questions or comments appreciated. Aug 18 22:49:23 Huh, so it unzips that tar as anticipated... so it's trying to do something else after the fact. Aug 18 22:57:29 halstead: ready for a master-next whenever is convenient Aug 18 22:58:06 RP, Should I queue the master-next first or try to make working 2.3.1 eSDKs first? Aug 18 22:59:27 halstead: you mean due to the missing sstate artefacts? Aug 18 23:00:05 halstead: I think I need to read mail Aug 18 23:00:21 RP, It's due to a non-routable IP address getting embedded in the eSDK making them nearly useless. Aug 18 23:00:39 RP, But 2.3 still works. Aug 18 23:01:18 halstead: right, 2.3.1 needs to be rebuilt Aug 18 23:01:29 halstead: do that first, then we can worry about -next Aug 18 23:01:37 Great. And I'll queue a next after. Aug 18 23:01:45 halstead: thanks Aug 18 23:01:52 RP, The "old" cluster is ready for a build too. Aug 18 23:01:59 We could start a master-next there. Aug 18 23:05:22 halstead: that works, lets try -next there Aug 18 23:05:37 RP, started. Thanks! Aug 18 23:06:55 halstead: thank you! :) Aug 18 23:07:40 halstead: it failed with the random oe-core repo issue we keep seeing :( Aug 18 23:07:58 halstead: nobody understands why that happens and it somehow fixes itself :/ Aug 18 23:08:09 I'll try again... Aug 18 23:08:43 Thanks for checking I wouldn't have noticed quite so fast. Aug 18 23:11:09 halstead: looks better this time :) Aug 18 23:12:04 RP is there a build cluster free this weekend? Aug 18 23:12:32 armpit: there should be some time after these builds Aug 18 23:12:42 k, i will monitor Aug 18 23:12:47 have a good weekend Aug 18 23:13:03 armpit: thanks, I think I'll be around dealing with M3 ;-) Aug 18 23:39:21 RP, Does https://autobuilder.yocto.io/builders/nightly/builds/480/steps/BuildImages/logs/stdio mean I need to re-queue? Aug 18 23:42:13 Seems like it's going fine. Aug 18 23:47:25 RP, Looks like I missed setting branch to pyro for meta-mingw. Probably need to cancel and restart. Aug 19 00:37:31 howdy Aug 19 00:37:40 just got made aware of this project Aug 19 00:37:49 it is of a similar purpose as a tool i am building Aug 19 02:51:14 Has anyone seen sysroots-uninative/x86_64-linux/usr/lib/libstdc++.so.6: version `GLIBCXX_3.4.23' not found (required by work/qemux86-poky-linux/core-image-sato-sdk/1.0-r0/recipe-sysroot-native/usr/bin/../lib/libapt-pkg.so.5.0) ? **** ENDING LOGGING AT Sat Aug 19 03:00:00 2017