**** BEGIN LOGGING AT Mon Sep 14 02:59:58 2015 Sep 14 05:50:38 hi guys Sep 14 06:29:21 morning all Sep 14 07:02:08 does anyone know how to pre-configure 2 network interfaces with the same IP address ? in order to have a proper network configuration whatever which port is actually up Sep 14 07:43:53 good morning Sep 14 08:10:08 Hello, anyone awake here? Sep 14 08:10:44 I am new to the yocto project, can you help suggest ebooks to learn and practice Sep 14 08:52:08 morning all Sep 14 09:00:03 hi bluelightning Sep 14 09:02:01 hi abelal Sep 14 09:20:39 Hello, I'm using binary jre from oracle in yocto x86_64 build, but it is not working... the java executable looks for ld-linux-x86-64.so.2 in /lib64. Sep 14 09:21:40 what's the best practice to create a symlink from /lib/ld-linux-x86-64.so.2 to /lib64/ld-linux-x86-64.so.2 ? Is there a way in image recipe to make the changes after creating the rootfs ? Sep 14 09:30:56 Does somebody knows how to make a symlink from /lib/ld-linux-x86-64.so.2 to /lib64/ld-linux-x86-64.so.2 after the image's do_roofs ? I've some incompatibilities with some prebuilt software looking in /lib64. Sep 14 09:45:17 Piziwate: you could do it by adding a function that you then call from ROOTFS_POSTPROCESS_COMMAND Sep 14 09:45:25 (in the image recipe) Sep 14 10:59:58 Ok good idea ! Thank you ! Sep 14 12:36:10 hell Sep 14 12:36:34 hell Sep 14 12:36:37 hello Sep 14 12:36:56 I am new to the yocto project, can anyone help me get started Sep 14 12:37:38 * dreamer88 slaps ccaione around a bit with a large fishbot Sep 14 12:38:19 * JaMa slaps dreamer88 around a bit with https://www.yoctoproject.org/documentation Sep 14 12:40:31 dreamer88: especially the quick start. Sep 14 12:43:41 slightly different topic: on dizzy, we're experiencing that an image build continuously pulls in an older uImage, although the recipe was changed, and the dependency tracking correctly triggered a rebuild. any pointers on where to start hunting? Sep 14 12:49:46 additional note: clansstate on the image fixed it. but why? :-) Sep 14 12:58:47 Hi there, Sep 14 12:58:48 i just wanted to asked if anybody was ever successful building cef3 recipe (chromium embedded framework) on a freescale IMX6. Sep 14 14:28:51 Hi! Sep 14 14:29:03 I've been fiddling with nios2 recently Sep 14 14:29:18 I'm trying to build core-image-minimal, but I'm running into issues with ptest Sep 14 14:29:38 more precisely, the tests are installed into the build/ directory instead of build/.libs for some reason Sep 14 14:29:51 this happens at least for dbus-test and libxml2 packages Sep 14 14:30:03 I wonder if there's something wrong with libtool Sep 14 14:32:02 with binaries, libtool generally writes binaries to .libs and then wrapper scripts for all those binaries in the parent above .libs. so files of that name should exist in both paths, just one is a wrapper script. Sep 14 14:34:01 kergoth: oh Sep 14 14:34:17 kergoth: so why would libtool put the binaries directly into the dir above .libs ? Sep 14 14:34:38 kergoth: I thought LT_OBJDIR might be the problem, but that's not the case Sep 14 14:55:29 hi Sep 14 14:56:00 simple question can I use regexp in @base_contains function in place of second param? Sep 14 14:56:26 like @base_contains('MACHINE', 'mach1|mach2' ... Sep 14 14:57:53 marek__: you want bb.utils.contains_any() instead Sep 14 14:59:23 bluelightning: many thanks Sep 14 15:12:12 kergoth: so either libtool on nios2 doesn't generate those wrappers OR it overwrites the wrappers with the binaries by placing them in the same dir Sep 14 15:16:20 usually when libtool doesn't know anything about an architecture, it does static linking. when things are statically linked then no wrappers are generated.. Sep 14 15:17:02 adding nios2 to OE/YP will require ensuring the architecture is known, not only to the build system, but to the various arch specific tools -- like libtool, autoconf, etc.. Sep 14 15:17:16 hmm what's that other bash script checking tool that isn't checkbashisms Sep 14 15:17:17 (usually pretty simply modifications in my experience) Sep 14 15:17:31 rburton, I usually just ask seebs to review it.. ;) Sep 14 15:17:38 ha Sep 14 15:17:51 we need SaaS, SeebsAsAService Sep 14 15:18:16 This is a perfectly reasonable idea and nothing could go wrong with it. Sep 14 15:18:20 http://www.amazon.com/Beginning-Portable-Shell-Scripting-Professional/dp/1430210435/ref=sr_1_5?ie=UTF8&qid=1442243877&sr=8-5&keywords=Peter+Seebach Sep 14 15:18:45 seebs: one instance could maintain pseudo, the other could check sh scripts for bashisms, the other could actually do your day job, and the other can relax Sep 14 15:19:09 rburton: shellcheck, http://www.shellcheck.net/ Sep 14 15:19:56 * kergoth yawns Sep 14 15:20:43 ah yeah if i change my hashbang to sh that produces 9 comments Sep 14 15:20:51 huh $"…" is a new one to me Sep 14 15:21:15 fray: the binaries are dynamically linked though Sep 14 15:21:35 fray: also, binutils, libtool and friends already do know about nios2 Sep 14 15:22:48 then libtool should be generating the wrappers... as Kergoth mentioned, the purpose of the wrapper is to make sure the exectuable that was linked can find the libs Sep 14 15:22:55 (until you do the make install) Sep 14 15:22:58 kergoth: "download and run locally!" it says. haskell! pah. Sep 14 15:23:28 fray: which is why I suspect it overwrites the wrappers with the dynamically linked binary Sep 14 15:23:49 fray: but why would that happen, that's a mystery to me Sep 14 15:46:51 rburton: heh, it's not too bad. install ghc, install cabal-install, run cabal install shellcheck, or https://github.com/kergoth/dotfiles/blob/master/scripts/cabalv will install it into a dedicated sandbox for that one app and symlink it to ${XDG_DATA_HOME:-~/.local/share}/../bin Sep 14 15:47:03 presumably some distros package it too, of course Sep 14 15:47:55 kergoth: turns out debian packages it Sep 14 15:48:04 ah, nice Sep 14 15:49:20 * kergoth has a pile of these shell scripts for per-app environments in his homedir. pip/pip3, cabal, cpan/perl, gem/ruby, brew (homebrew), npm, lua Sep 14 16:04:45 hi guys Sep 14 18:30:36 Hmm, do we have a per-layer way to silence the 'no recipe files match the BBFILE_PATTERN' warning? Sep 14 18:31:09 this layer has appends, but they're layer specific, and those layers aren't always included, so in some configurations there won't be any, but in others there will Sep 14 18:31:18 i hate to add a no-op empty bbappend just to shut it up, but if i have to.. Sep 14 18:33:14 ah, we do, nice Sep 14 18:33:35 BBFILE_PATTERN_IGNORE_EMPTY_ Sep 14 19:17:42 * nerdboy looking at the 1.8 dev manual Sep 14 20:12:31 Anyone else noticed that changing IMAGE_CMD_ doesn't cause do_rootfs to rerun? Sep 14 20:25:55 RP: ping Sep 14 20:26:01 or bluelightning_ perhaps Sep 14 20:27:05 I think that might be one of those situations where it can't figure out the dependencies by itself Sep 14 20:27:21 I hadn't noticed that one specifically though Sep 14 20:28:00 we do explicitly handle IMAGE_TYPEDEP_, but not IMAGE_CMD_, whcih seemed odd Sep 14 20:28:02 will look more closely Sep 14 20:28:21 on another note, do you happen to know if forking bitbake would interfere with yocto-compliance? Sep 14 20:28:35 iirc there's a line about everything having been pushed upstream in the terms Sep 14 20:28:51 (we might need my git shallow support for the next mel release, and it won't be merged in 2.0) Sep 14 20:31:23 * kergoth checks the site, assuming he can find the link again Sep 14 20:33:44 ah, looks like patches have to have been submitted, but it doesn't say they have to have been merged :) Sep 14 20:36:45 gah, the IMAGE_TYPEDEP vardep is wrong too Sep 14 20:36:59 'IMAGE_TYPEDEP_' is added to vardeps directly, without the IMAGE_FSTYPES-based suffix Sep 14 20:45:44 kergoth: yes, patches just have to be shared, not merged Sep 14 20:46:08 kergoth: not noticed IMAGE_CMD but it doesn't surprise me :( Sep 14 20:48:57 RP: k, thanks Sep 14 20:50:01 RP: I've been thinking about getting a bit more active with yocto outside of work, though how much time i'll have i don't know, but let me know if you think of anything I can do to help reduce your workload. at a minimum I'm going to try to do more bitbake patch review ack'ing Sep 14 21:00:38 kergoth: that would be good thanks! I think in the main part I'm really struggling with are the intermittent "random" autobuilder failures we keep seeing. If you fancy picking any of those up you'd be welcome. Other ideas are some of the less loved parts of the system like systemd, or some of the general code "infrastructure" improvements we keep talking about Sep 14 21:01:34 okay, I'll take a look at the autobuilder first Sep 14 21:04:45 kergoth: the ones that are still open mostly involve systemd at this point, gettys timing out and such Sep 14 21:04:53 lovely Sep 14 21:05:11 kergoth: they've in the bugzilla. I know the AB can have high load but I can't shake the feeling that there is some underlying in systemd Sep 14 21:05:39 kergoth: I think putting sleeps in the qemu io thread was able to replicate some of it Sep 14 21:06:23 kergoth: We've just taken a significant systemd version upgrade so its at least possible that might help. I only merged it a few hours ago so too soon to tell though. Sep 14 21:46:14 kergoth: thinking more about that, also try and find something you'll enjoy... Sep 14 23:47:26 this libtool is really horrible, 11k lines of insane shell script Sep 14 23:47:55 i recommend no one ever read libtool Sep 14 23:48:00 but if you do, don't read the generated script Sep 14 23:48:13 start with ltmain.sh, but iirc even that is partially generated by macros in order to avoid use of shell functions Sep 14 23:48:17 kergoth: well, I have trouble with libtool, so how am I supposed to debug that ? Sep 14 23:48:39 assuming the problem occurs outside of oe as well as in, email the libtool list for advice Sep 14 23:50:10 kergoth: that might be the best approach now I guess Sep 14 23:50:34 kergoth: I have to wonder why libs are placed into .libs , but the wrappers are placed one level lower and then immediatelly overwritten by the binaries Sep 14 23:51:58 have you verified that's exactly whats happening? remake might be of use to debug what the makefile is running Sep 14 23:53:14 remake ? what's that ? Sep 15 00:00:17 http://bashdb.sourceforge.net/remake/ Sep 15 00:00:39 a version of gnu make with improved tracing and debugging Sep 15 00:00:54 oh nice Sep 15 00:02:22 make -d is pretty much useless, so it's nice to have an alternative to see what's going on Sep 15 00:02:24 heh Sep 15 00:04:59 kergoth: thanks Sep 15 00:05:11 kergoth: I think I'll hit the hay for tonight, I feel somewhat crappy Sep 15 00:05:20 np. night **** ENDING LOGGING AT Tue Sep 15 02:59:58 2015