**** BEGIN LOGGING AT Mon Dec 04 03:00:02 2017 Dec 04 09:37:35 hi, does exist in rocko some variable which points to recipe sysroot ? Dec 04 09:41:04 STAGING_DIR_[native|host|target] ? Dec 04 10:19:48 nayfe: ok thanks I'll try Dec 04 10:20:30 nayfe: but it not points to the recipe sysroots Dec 04 10:21:58 naybe: but to ${TMPDIR}/sysroots which is in case of rocko empty Dec 04 10:23:05 open-nandra: what is your recipe ? Dec 04 10:23:23 nayfe: have custom application Dec 04 10:23:32 I'm using cmake to build it Dec 04 10:24:27 nayfe: I need to pass somehow recipe sysroots Dec 04 10:25:15 if you do something like bitbake -e | egrep '^[A-Z][A-Z_]*' | grep recipe-sysroot Dec 04 10:26:02 to me STAGING_DIR_TARGET is pointing to it Dec 04 10:27:22 nayfe: stupid me I use only STAGING_DIR (without target) now everything work as expected Dec 04 10:27:28 nayfe: thanks again ;) Dec 04 10:27:44 open-nandra: np :) Dec 04 10:33:36 if you need to do *anything* in a recipe apart from inherit cmake then you're doing it wrong Dec 04 10:39:53 rburton: sometimes, cmake projects are using custom variables :) Dec 04 10:40:25 in which case they're doing it wrong ;) custom application implies open-nandra wrote the cmake Dec 04 10:41:42 rburton: indeed, and it leads to massive patches :p but indeed, in open-nandra case, it can be solved in cmake stuff Dec 04 10:42:54 indeed Dec 04 10:43:01 never 2 without 3 Dec 04 10:44:29 ouch english ppl dont say that Dec 04 10:45:07 * nayfe plugs brain on Dec 04 10:51:59 i was wondering what you were talking about ;) Dec 04 10:54:58 rburton: lol :) i just said 2 "indeed" in first sentance, so I put a third "indeed" as all things comes in three (in french the equivalent expression is "never 2 without 3") Dec 04 10:55:11 hard monday Dec 04 10:55:40 ah right, interesting Dec 04 10:56:02 we have "things come in threes" but there's no need to do a third if you do something twice Dec 04 10:56:27 can we call that my education of the day and i can go watch netflix instead now? Dec 04 10:56:45 sure! Dec 04 11:44:00 rburton: is dnf/rpm upgrade dropped again? Dec 04 11:44:15 haven't queued it again yet Dec 04 11:44:36 i'll pull it into the new mut as RP merged the last one over the weekend Dec 04 11:44:57 sure, it's a bit hard to deduce why some patches are quicker to land than the others :) Dec 04 11:45:09 becayse they're easier and/or don't get implicated Dec 04 11:45:18 is there a branch with the outstanding patches in? Dec 04 11:45:31 rburton: on my local machine; I can push it Dec 04 11:45:38 please do and rebase to current master Dec 04 11:45:44 rburton: oh, and meson ;) Dec 04 11:45:51 yeah thats a biggie Dec 04 11:48:08 rburton: actually, I need to first check if upstream has released something newer meanwhile, and fix the psmisc translations issue, and merge the rpm updates and all the other updates into one branch, and do a smoke test... and I have a sore throat today :( Dec 04 11:48:17 so not today perhaps? Dec 04 11:48:36 fine by me Dec 04 11:49:14 RP: so f26 is falling over in qemu again i see Dec 04 11:49:23 we can we just reboot it every night Dec 04 11:49:40 * rburton wishes he wasn't only half joking Dec 04 12:07:47 New news from stackoverflow: Kernel selection using yocto Dec 04 12:26:02 rburton: it is? memory alloc or the hang? Dec 04 12:26:33 hang in boot this time mostly Dec 04 13:33:24 RP: are you rejecting my unicode progress bar patch on purpose? Dec 04 13:33:56 ██████ > ###### Dec 04 13:54:49 RP: huh, master selftest failed with sametune failure for initscripts. i saw that once, but it disappeared forevermore and never appeared on the ab since. Dec 04 14:08:11 New news from stackoverflow: menuconfig window pop-up issue in linux Dec 04 14:13:07 rburton: I'm nervous about that patch Dec 04 14:19:43 RP: we mandate a utf8 locale and pretty much everyone has non-shit fonts Dec 04 14:32:56 rburton: you didn't see the mess my irc client made of that above :) Dec 04 14:33:19 well thats your fault ;) Dec 04 14:34:19 irc clients are special, iirc irc doesn't have an encoding on the server Dec 04 14:38:14 New news from stackoverflow: how to add the binaries and other files to rootfs in yocto using .bb recipes Dec 04 14:41:50 rburton: I guess I quite like old school hashes ('pounds') too Dec 04 16:58:07 RP, sent v2 rocko-pull request. Dec 04 17:10:02 armpit: does that fix the oe-selftest issue? Dec 04 17:10:38 yes Dec 04 17:11:39 the tip of stable/rocko-next has the inode work around patch. i did not include it in the pull request Dec 04 17:12:23 want me to remove it ? Dec 04 17:12:55 armpit: its ok, I can avoid it but thanks for the warning Dec 04 17:13:11 rburton: finally we get progress bars as good looking as DOS had in 1981 ;) Dec 04 17:13:38 Do people dislike '#' that much? :) Dec 04 17:14:26 RP: I'm just feeling nostalgic for DOS' far superior pseudographics, no hard feelings towards # Dec 04 17:15:47 has anyone had an error "contains proably-redundant RPATH /usr/lib [useless-rpaths] Dec 04 17:24:35 jdelmore2: I once fixed a lot of them... Dec 04 17:24:39 kanavin1: fair enough :) Dec 04 17:24:43 armpit: merged, thanks! Dec 04 17:24:56 armpit: do we have a fontconfig patch for morty I can pull in? Dec 04 17:26:00 yes. plus more. http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/log/?h=akuster/morty-next Dec 04 17:26:14 armpit: what is the build status? Dec 04 17:26:41 one of the tests are still failing.. Dec 04 17:26:46 * armpit looks Dec 04 17:26:46 armpit: you mean stable/morty-next? Dec 04 17:26:59 ah, np Dec 04 17:27:01 no Dec 04 17:31:00 RP: I didn't know that Unicode actually includes a full replica of it :) https://en.wikipedia.org/wiki/Block_Elements and a superset of line drawing characters too Dec 04 17:31:12 RP, sorry. I have not convert morty to stable/ format.. let me fire off a build Dec 04 17:37:54 armpit: thanks. I've pulled the fontconfig one in since the morty build clearly need ti now Dec 04 17:39:08 k Dec 04 18:44:55 Hi, I am trying to build an image for the Digi ConnectCore 6UL. I run bitbake core-image-base-swu and it executes fine, but the resulting image is not updated in tmp/deploy/image/ccimx6ulstarter. I guess it has something to do with the cache, how can I force an image re-generation? Dec 04 18:45:42 And also, how can I see the content of an image's rootfs? Dec 04 19:17:05 nobody? Dec 04 19:19:10 alright thanks anyways Dec 04 19:19:24 bye Dec 04 19:31:04 @RP is there a quick fix for teh RPATH problem? Dec 04 20:49:14 im having trouble with a recipe. I specify S = "${WORKDIR}/git", SRC_URI = "git://blabla.bit file://anotherblabla.patch", log shows patch being applied, work directory and running the binary tells me it was not Dec 04 20:49:45 also SRCREV = "commit hash from the git repo" Dec 04 20:52:06 read log.do_patch to see what it patched. that *should* work. Dec 04 20:52:18 (and obviously does for all the recipes in oe-core which do that) Dec 04 20:52:54 fun trick: change the patch to introduce a syntax error and verify it breaks the build Dec 04 20:57:57 i've looked at log.do_patch and see a note saying "applying patch bla bla bla" Dec 04 20:58:23 however it seems to not actually apply my patch Dec 04 21:01:47 uglyoldbob: I assume git status shows no change? Dec 04 21:02:23 uglyoldbob: also, you may want to try a bitbake -c clean && bitbake -c patch Dec 04 21:02:52 The clean task will wipe your directory clean though, so make sure your mods are out of WORKDIR first Dec 04 21:05:58 overdamped: no change, git status in the tmp/work/blabla/recipe/git shows patches/series has changed Dec 04 21:06:46 There's nothing odd in your log.do_patch? Dec 04 21:08:17 overdamped: nope, everything indicates success Dec 04 21:09:21 New news from stackoverflow: Creating Yocto Mender for Raspberrypi 3 Dec 04 21:18:13 any other ideas? ive tried deleting work/tmp and rebuilding, no change Dec 04 21:23:24 uglyoldbob: my only suggestion is what its patching isn't what you expect Dec 04 21:25:59 jdelmore2: depends where its coming from. Its usually a question of figuring out what added it and why, then figuring out how to stop that. Often its due to libtool not using our macros but some which ship with the source Dec 04 21:39:43 so for some reason i have to force the patch step to run in order to get it to work correctly Dec 04 21:52:00 using devtool to make some kernel modifications. Trying to add a kconfig item results in config error pastebin.com/ZpTbnqAJ Dec 04 21:53:49 how do I force devtool to rebuild config? Dec 04 21:54:09 nathani__: can you let me know what exact steps got you to this point? Dec 04 21:55:19 bitbake -c menuconfig linux-boundary ? Dec 04 21:56:52 installed sdk->setup enviro->devtool build-image my-image->devtool modify virtual/kernel->make edits->devtool build-image my-image Dec 04 21:57:37 uglyoldbob, don't think I have access to bitbake in SDK Dec 04 21:58:57 shoot bluelightning, i apologize, i posted wrong error. This was after i deleted .config hoping that would force a rebuild Dec 04 22:00:05 pastebin.com/TYfK3ew9 Dec 04 22:04:38 I did try running make oldconfig, but that was pretty horrible. Dec 04 22:05:42 my apparently broken recipe is here: https://github.com/uglyoldbob/oe-custom/blob/verdex/recipes-bsp/u-boot/u-boot-verdex_1.2.0.bb Dec 04 22:08:37 uglyoldbob, does that work for filesextrapaths? PN-PV? Dec 04 22:09:24 your folder is "u-boot-verdex" Dec 04 22:09:33 New news from stackoverflow: Is it normal multiconfig is very slow to parse? Dec 04 22:15:14 I usually go for a ${THISDIR}/files for simplicity Dec 04 22:16:22 you should not be setting filesextrapaths in a recipe at all, it's pointless. PN-PV is in the default search path for any recipe already Dec 04 22:16:40 the whole reason FILESEXTRAPATHS exists is for bbappends to easily add to that search path relative to the bbappend path, which isn't in the default search path Dec 04 22:16:55 that said, setting it should be harmless Dec 04 22:26:40 kergoth, good point Dec 04 22:27:55 i suppose some added information for my issue is that I edited a file in oe-local-files to attempt to set my new kconfig item. this is probably the cause of my issue I would think. Dec 04 22:39:39 New news from stackoverflow: Bitbake's order of execution of Python-based run files? Dec 04 23:37:55 halstead: is http://downloads.yoctoproject.org/mirror/sources/ normally super slow? Dec 04 23:40:44 clsulliv: Not normally. It's a 4gbps link. Where are you downloading to? Dec 04 23:41:48 halstead: just trying to view in a web browser atm Dec 04 23:42:12 clsulliv: At Intel via the proxy? Dec 04 23:42:38 halstead: yes. The parent directory loads instantly, but that one seems to time out Dec 04 23:43:44 clsulliv: indexes should be disabled there. Indexing that directory puts a huge load on nfs. I thought it would return "forbidden" when trying to grab an index. Dec 04 23:44:32 clsulliv: Is this something that you need to see web broswer indexes? I maybe ought to generate a static index file every few minutes. Dec 04 23:46:26 halstead: bavery and I were seeing a fetcher failure earlier, and bavery noticed that the mirror contained the previous version of the package but not current, I was just going to take a look myself via the index because he didn't give a full link. Dec 04 23:47:59 halstead: at one point the page did come up for me, but it didn't look complete. Contained a number of patches and what looked like the autobuilder hierarchy rather than sources Dec 04 23:49:04 halstead: e.g., downloads.yoctoproject.org/mirror/sources/home/pokybuild/ Dec 04 23:49:50 There are some timeouts happening. Trying to correct now. Dec 04 23:53:07 clsulliv, Yeah, i don't know why there is a home/pokybuild dir in sources. That seems wrong. Dec 05 00:13:11 FILESEXTRAPATHS is handy for consolidating patches for python2 python3 recipes and not repeating yourself repeating yourself **** ENDING LOGGING AT Tue Dec 05 03:00:01 2017