**** BEGIN LOGGING AT Fri Sep 29 03:00:01 2017 Sep 29 03:04:58 User__: what I just mentioned above, you add it on SRC_URI on your kernel recipe file and copy it to the directory Sep 29 03:05:15 User__: on the kernel dev manual says how to do it http://www.yoctoproject.org/docs/latest/kernel-dev/kernel-dev.html\ Sep 29 03:07:56 aehs29, Ok. thanks. Just was not sure if what you said before was optional Sep 29 04:52:49 https://eula-downloads.yoctoproject.org/index.php Sep 29 07:49:27 Hello has any of you experieced the SSL: CERTIFICATE_VERIFY_FAILED on a yocto built image? Sep 29 07:51:25 I have successfully included openssl and PACKAGECONFIG_append_pn-openssl= " perl" to enable the c_rehash function, but there is no sign of the c_rehash function... Sep 29 07:51:39 I believe the problem lies there... Sep 29 08:18:35 Good morning, Sep 29 08:19:05 If I may ask - is there any description/documentation about the OE build system caching ? Sep 29 08:22:10 the sstate-cache magic ;) Sep 29 08:27:56 nayfe: yes :), exactly Sep 29 08:28:04 The arcane..... part of OE Sep 29 08:28:22 (and when some strange errors show up - remove caches :D ) Sep 29 08:28:25 Hey, is there a solution to layer multiple overlayfs? Sep 29 09:45:47 Is hob commonly used to build with poky? Sep 29 09:46:51 not really Sep 29 09:47:04 it was never really finished, had bugs, and has been removed for quite some time Sep 29 09:52:28 I see Sep 29 09:53:24 I'm reading open embedded linux with yocto project and there is mentioning of hob Sep 29 09:54:15 sorry, embedded linux development Sep 29 10:12:55 lukma: sstate-cache is not perfectly working, depends on your scenario Sep 29 10:14:59 User__: have a look at Toaster Sep 29 10:16:03 mckoan, Will I be getting toasted? Sep 29 10:18:06 User__: :-) Sep 29 10:19:26 mckoan: I would like to have better understanding of it Sep 29 10:19:36 hence I'm looking for some documents to read Sep 29 10:19:47 (probably for some non OE expert) Sep 29 10:20:50 mckoan, I would like my image black please, both sides. Sep 29 10:21:35 FabKna: did you read https://www.kernel.org/doc/Documentation/filesystems/overlayfs.txt ? Sep 29 10:21:46 mckoan, while the kernel still moist.. Sep 29 10:23:16 has there been a change lately to the yocto ML in terms of moderation? Sep 29 10:23:23 lukma: https://wiki.yoctoproject.org/wiki/Enable_sstate_cache#Use_local_sstate_cache Sep 29 10:26:57 mckoan: thx Sep 29 10:26:58 :) Sep 29 10:31:51 meh, no moderation change, local postmaster tricked me. Sep 29 10:32:03 * LetoThe2nd manages to look like an idiot on the ML. hooray! Sep 29 10:43:23 hello Sep 29 10:43:40 is there a way to tell fetch() to produce debug output? Sep 29 10:43:49 would -DDD do this ? Sep 29 11:00:56 yes, in the logs Sep 29 11:45:48 thank you rburton Sep 29 12:19:28 rburton: soooooooooooooo.... Sep 29 12:19:44 rburton: all of this rocket science to change gcc and binutils around to use a different target prefix Sep 29 12:20:03 it doesn't actually work very well, because everything else is still in the old prefix... ;) Sep 29 12:20:18 so gcc can't find libgcc. immediate fail. Sep 29 12:22:02 I'm actually wondering if it wouldn't have been quicker to just smash TARGET_SYS to the value I wanted and mop up the damage after all Sep 29 12:47:04 I want multiple partitions in my sdcard image. How to do this? Sep 29 13:30:35 ramcq: yeah you might be right :/ Sep 29 13:31:18 :'( Sep 29 13:50:06 Is Ed Bartosh around? Sep 29 13:55:15 FabKna: with wic, you can specify multiple partitions Sep 29 13:55:44 example here: https://community.nxp.com/thread/389816 Sep 29 14:10:07 hello, I have run into a strange error after trying a new machine configuration file instead of the vendor one: https://pastebin.com/1Tb7Bmd1 Sep 29 14:10:23 how do I even start debugging this? Sep 29 14:11:59 this is my machine configuration: https://pastebin.com/jVRD9sU1 Sep 29 14:39:58 eduardas_m: looks like you are running on an existing build for a different machine, try the simpler version is you're advised to wipe out tmp and rebuild Sep 29 14:54:03 nayfe: Ty! Sep 29 14:56:41 hmm, my linux-xlnx build fails with "permission denied" while trying to run `fixdep` Sep 29 15:01:13 armpit: around? I am looking at the morty lsb failure and want to see the AB log itself, do you have a pointer to that one? Sep 29 15:15:53 mckoan: thank you for the advice, but even wiping tmp and sstate-cache did not help Sep 29 15:16:03 got same error Sep 29 15:17:17 eduardas_m: ti-compat-wireless and fod-som-linyx-variscite both have the same files and your build has them both in Sep 29 15:17:30 eduardas_m: you need to decide what one you want and how to stop the other being pulled in Sep 29 15:24:55 rburton: is it because in linux-variscite recipe there is only COMPATIBLE_MACHINE = "(var-som-mx6|imx6ul-var-dart|imx7-var-som)" and that does not include my machine fod-som? Sep 29 15:26:42 I get it that my machine includes ti-compat-wireless via MACHINE_EXTRA_RDEPENDS, however I can not understand where from the same files come Sep 29 15:27:22 how do I know what is the second recipe that wants to include the same files? Sep 29 15:27:44 the message tells you Sep 29 15:28:19 ti-compat-wireless-wl18xx and fod-som-linux-variscite Sep 29 15:30:19 rburton: I somehow missed that, thank you Sep 29 15:40:57 rburton: thus far every package I've found with ARM_INSTRUCTION_SET="arm" (to disable thumb) has built OK with thumb turned back on (like every Linux distro does for armhf) Sep 29 15:41:13 rburton: I think some of this might be from like, gcc 2.9 on armv4 in 2004 or something Sep 29 15:44:34 ramcq: some maybe, some were removed and re-added recently Sep 29 15:50:25 Does Yocto produce any config files as part of the target image that can be viewed? For example, if I wanted to see how it was config'd on a running system, is any of that info there? Or is it only available on the host machine that produced the image? Sep 29 15:52:24 clevernickname: nothing like that goes onto the target by default, no Sep 29 15:56:27 did bitbake's order of operations change recently? I'm adding to BB_HASHBASE_WHITELIST in a ConfigParsed handler. this used to work, the vars there were set by the time SanityCheck ran, but that's not the case anymore Sep 29 15:56:28 rburton: Thanks. I know the kernel has the option of including .config... and was wondering if Yocto did anything similar. Sep 29 15:58:00 oh, wait, i see, nevermind Sep 29 15:58:01 i'm blind Sep 29 15:58:24 * kergoth rolls eyes at self, obviously it's BB_ENV_EXTRAWHITE i need. insufficient caffeine Sep 29 16:00:12 clevernickname: iirc we don't ship the config in a package but it should be trivial to do if its not already an option Sep 29 16:44:41 halstead: are you working on the ABs today? Sep 29 16:51:36 halstead: dinner is ready so i'm off. if you need to shut down the autobuilder can you restart nightly against ross/rocko and ross/mut please :) Sep 29 17:15:08 Sorry I missed you rburton. I snuck in an did updates on Wednesday. There were some urgent ones. So I won't need to interrupt anything. Sep 29 17:18:39 ah, this fc-cache hang only happens on x86_64. wonder if its a qemu or fontconfig bug Sep 29 17:18:52 anyone else try a qemux86-64 build with fontconfig installed in an image recently? Sep 29 17:52:43 kergoth: ab, often. Sep 29 17:54:31 i'm hitting the hang every time, poky distro, qemux86-64, only meta-sourcery added. digging further Sep 29 17:56:05 actual hang? Sep 29 17:57:54 yep, the qemu-x86_64 fc-cache process has been running for an hour, and interrupting the bitbake leaves that process behind. i have a handful i had to kill -9 Sep 29 17:58:04 * kergoth removes meta-sourcery and meta-mentor-staging to try fully stock Sep 29 17:58:20 I'm having trouble with a python recipe I've made. The dependencies are python-future and python-lxml. Both build just fine. However when building my recipe setup.py fails with "No module named future". Even though python-future is in RDEPENDS and has been build with no problems. Sep 29 17:59:32 It's like the path hasn't been added or something. Am I missing something? Sep 29 18:00:02 well build depends should be DEPENDS not RDEPENDS Sep 29 18:00:33 and if you want the native python to see them to run the build tools, you'll want to depend on the native variants Sep 29 18:00:38 and if its trying to *run* code at build time using future then you need to inherit pythonnative and DEPEND on python-future Sep 29 18:00:43 erm, python-future-native Sep 29 18:00:43 :) Sep 29 18:01:46 huh, my good for rocko rc2 branch is failing rather badly Sep 29 18:01:59 halstead: WARNING: The free inode of /srv/autobuilder/autobuilder.yoctoproject.org/pub/sstate (systemd-1) is running low (58.453K left) Sep 29 18:02:11 armpit: around? Sep 29 18:02:23 phew its the ab dying! Sep 29 18:02:58 halstead: really off now, can you restart those nightlies when the ab is back? Sep 29 18:03:07 rburton: thanks a bunch Sep 29 18:03:17 rburton, Sure. Sep 29 18:14:01 rburton, I'm not sure what's up since there are 4177271K inodes left on that volume not 53K. Looking into the disk space monitor now. Sep 29 18:31:25 overflow problem? Sep 29 18:34:59 fray, Perhaps. We've had this many inodes available before with no issues. Maybe the used value (47754741) is overflowing? It has certainly grown. Sep 29 18:40:59 hey everyone. is there a structure for multi-layer yocto projects you recommend? Sep 29 18:41:44 my ideas so far are git submodules, the google `repo` command and that's about it^^ Sep 29 18:46:05 peacememories: both good choices Sep 29 18:46:41 it still feels kind of brittle in terms of keeping everything in sync Sep 29 18:48:40 how? submodules etc are designed to keep stuff in sync... Sep 29 18:48:58 just telling your users to grab random SHAs or track a specific branch manually would be brittle Sep 29 18:51:44 I try to create a not mounted partition with wic (wks) file. How to achieve that? Sep 29 18:52:07 also making sure that all layers land in bblayers.conf would probably require duplication Sep 29 18:52:15 not mounter is not correct. It should be empty. Sep 29 18:52:26 i'm not saying it's not workable, i just wish there was a clear "this is a good way to do this" method Sep 29 19:03:12 rburton: I don't suppose the DISTRO used for a Yocto build is put any where on the target either? Sep 29 19:04:04 a propos distro. i don't quite get what a distro actually is in yocto, and in which case i would want to create my own Sep 29 19:06:56 peacememories: I do NOT have firm grip on things myself, but I believe DISTRO == build policies for the image. Sep 29 19:28:08 peacememories: we don't mandate a method, you can pick. some use repo, some use submodules, some combo-layer. your choice. Sep 29 19:28:41 peacememories: yeah basically policy in the loosest possible sense. what machines, what image, what configuration, what recipes. Sep 29 19:29:20 peacememories: if you're using yocto for something non-trivial you should make a custom distro. its trivial. Sep 29 20:16:52 lovely. fc-cache hangs under qemu user emulation only when built by the sourcery toolchain. it runs fine under qemu system emulation, or when built by the oe-core toolchain Sep 29 20:17:00 fun fun Sep 29 20:49:13 hmm, are there any guides on creating a custom distro, (and when it makes sense)? googling seems to only give me results using "distro" in the classical sense (a la "create a custom distro with yocto" ^^) Sep 29 20:52:48 poky is a fine example. just copy meta-poky/conf/distro/poky.conf to your own layer with your own distro name and modify however you want Sep 29 20:52:56 or base on one of the many other distros, really your call Sep 29 20:53:08 or create a new one, and see defaultsetup.conf for vars that might be of interest Sep 29 20:53:23 a distro is just a config file that sets variables that control how things are built, what package manager is used, etc Sep 29 20:53:32 overarching policy, generally Sep 29 20:59:21 okay, thx Sep 29 21:07:52 I try to configure read only rootfs with overlayfs with https://github.com/pedwo/meta-readonly-rootfs-overlay . However, it does not work. Does somebody has a bit experience with that? **** ENDING LOGGING AT Sat Sep 30 03:00:01 2017