**** BEGIN LOGGING AT Mon Dec 19 03:00:00 2016 Dec 19 03:19:29 Hello all, Dec 19 03:22:09 Am building yocto for warp7imx machine. But the image that i build does not have some needed commands like modprobe.is there a general package that can be added in. Local.cong Dec 19 03:23:16 ...in local.conf file via image_install or something? Dec 19 03:23:53 AndroUser: the package providing modprobe etc would be "kmod" I believe Dec 19 03:25:00 Oh Dec 19 03:27:58 Blueligjting: i have added i2c tools and tools-sdk via image_install_append and extra_image_features respectively. What other essential packages could be added to make it full fledged? Dec 19 03:33:02 . Dec 19 04:10:03 madhav: it depends entirely what you mean by "full fledged" Dec 19 04:11:23 madhav: perhaps you could add "packagegroup-core-full-cmdline" though, that will give you e.g. coreutils instead of what busybox provides in substitute Dec 19 04:13:31 Bluelighting:presently i have to add gdb,also i have to enable wifi and ble support Dec 19 04:13:46 Oh ok Dec 19 04:44:26 [10:14] -NickServ- This nickname is registered. Please choose a different nickname, or identify via /msg NickServ identify freenode Dec 19 15:54:45 I am trying to build fitImage images but battling to find documentation/examples, can anyone point me to some helpful resources please? Dec 19 17:07:51 hi, is anyone around whom I can ask some questions about a bunch of build problems I've been having? Dec 19 17:09:00 possibly, but be aware that IRC is pretty asynchronous. Dec 19 17:09:08 Usually just go ahead and post questions and then wait a few hours. Dec 19 17:13:24 @seebs, fair enough Dec 19 17:13:34 beh, IRC tags without having to use at's. Dec 19 17:16:19 basically, I'm building a custom Yocto image, and on top of loads and loads of host-user-contaminated errors (which I think are due to the way I'm running bitbake inside a Docker container), I'm getting this: http://pastebin.com/wGtejpZ2 Dec 19 17:16:39 "ERROR: e2fsprogs-native-1.43-r1 do_populate_sysroot_setscene: Error executing a python function in exec_python_func() autogenerated" Dec 19 17:19:26 [23:26:53]W: [Step 1/1] Exception: struct.error: unpack_from requires a buffer of at least 4 bytes Dec 19 17:19:57 That's all part of the uninative bbclass that's built into Poky these days, but I don't really understand why the error is happening. Dec 19 17:30:44 When I comment the following out, my build works: Dec 19 17:30:44 #echo 'IMAGE_GEN_DEBUGFS = "1"' >> conf/local.conf #echo 'IMAGE_FSTYPES_DEBUGFS = "tar.bz2"' >> conf/local.conf #echo 'IMAGE_INSTALL_append = " gdbserver"' >> conf/local.conf Dec 19 17:30:58 hold on, brb Dec 19 17:32:09 ok, back Dec 19 17:32:18 https://www.irccloud.com/pastebin/7UNYG9xE/ Dec 19 17:32:23 much better Dec 19 17:32:30 commenting out those three lines seems to let things work Dec 19 22:53:27 anyone else seeing python3 sysroot file duplication issues that started ~1-2 wks ago? Dec 19 22:53:49 /home/paul/poky/build/tmp/sysroots/x86_64-linux/usr/lib/python3.5/distutils/__pycache__/spawn.cpython-35.pyc Dec 19 22:53:49 Matched in b'' Dec 19 22:53:49 Please verify which recipe should provide the above files. Dec 19 22:54:01 and 58 other instances of similar. Dec 19 22:54:29 I'm re-using the same build dir day to day; not starting from scratch. Dec 19 22:54:58 if I cleansstate py3 and py3-native AND manually purge the sysroot of py3 then I'm good. Dec 19 22:55:13 if I omit the last step it still craps out. Dec 19 22:56:02 paulg: yes I have seen that Dec 19 22:56:42 I don't have any advice beyond what you did... in my case I just deleted tmp, which is exactly what I tell others not to do :/ Dec 19 22:56:44 bluelightning, ok, good (in a sense), so I can now say it isn't triggered by trash in my local layers. Dec 19 22:57:16 * paulg tries to avoid the nuke-and-pave approach of sending all of tmp to the bitbucket if possible. Dec 19 22:57:38 I've thought several times that we really need a regression test that detects when we break incremental builds with the same TMPDIR, because I suspect it happens more often than it should Dec 19 22:58:09 that "Matched in b''" is especially infuriating Dec 19 22:58:25 * paulg wants an "--if-pkg-fails-then-cleansstate-it-and-try-again" option. Dec 19 22:58:42 cleansstate wouldn't fix this though Dec 19 22:58:50 true, that doesn't work in this case. Dec 19 22:59:17 what it's telling you is something outside of sstate's (current) tracking has left files in the sysroot Dec 19 22:59:45 but I had a hardware issue a couple days ago where a machine crashed and since it was doing 16 pkgs at a time, it left 16 pkgs in a borked state. Dec 19 23:00:14 requiring a cleansstate and restart; rinse lather and repeat 16 times. Dec 19 23:00:41 yes, at that point I should have pushed tmp off a cliff ; on a 12 core xeon, it would have been faster. Dec 19 23:00:53 yeah that's hard to avoid - often it's outside of our control, if the system half-wrote an intermediate artifact that make now thinks is complete Dec 19 23:01:47 though perhaps we could have a shortcut to cleaning previously interrupted tasks (or mode where we do that automatically) Dec 19 23:02:16 in practice only certain tasks would suffer from that since others clean up as the first step (e.g. do_compile would have the issue, do_install would not) Dec 19 23:03:28 ya, 99% of the fails after the crash were in do_compile, complaining about .o and .a files being truncated/invalid ; no surprise. Dec 19 23:04:56 anyway, a corner case that most people don't care about -- the average user isn't doing daily builds looking for fallout from upstream changes. Dec 19 23:06:32 the initial issue we talked about is annoying to me at least Dec 19 23:12:15 ya, the py3 issue should be sorted one way or another ; I just know I'd burn a lot of hours trying to get a handle on it, and with a work-around, I've got other fish to fry. Dec 19 23:58:31 paulg: you are probably a victim of http://git.openembedded.org/openembedded-core/commit/?id=bb4d689769703177dbb0df0935e15016b879f42b Dec 20 00:14:51 khem, could be ; I'm generally python illiterate, so I can't look at that and say yes or no. Dec 20 00:21:36 paulg: try to revert it and see if it helps Dec 20 00:22:59 khem, I will do that next time I trigger it. Problem is once I've deployed my "workaround" I don't know what the trigger is that makes it re-appear. **** ENDING LOGGING AT Tue Dec 20 03:00:01 2016