**** BEGIN LOGGING AT Fri Jan 03 03:00:23 2020 Jan 03 04:59:57 New news from stackoverflow: How to source a Bash script in a Yocto recipe || Yocto tensorflow 2.0 Jan 03 06:17:22 When are the IRC channel logs of 2020 going to get updated? Jan 03 07:43:00 Autotools is installin python files in /home/testuser/work/project/build/tmp/work/aarch64-tdx-linux/pylib/3.3.0-r0/image/home/testuser/work/project/build/tmp/work/aarch64-tdx-linux/pylib/3.3.0-r0/recipe-sysroot/usr/lib/python3.5/site-packages/, what do I need to change to make it install correctly in ./image/usr/lib/python3.5/site-packages and not Jan 03 07:43:00 that whole /home/blablabla Jan 03 07:45:40 do I need to change something in the autoconf file or can I append/prepend something in the recipe? Jan 03 08:20:18 export PYTHON_SITE_PKG=${PYTHON_SITEPACKAGES_DIR} in do_configure_prepend seems to work Jan 03 09:57:24 what means error during do_rootfs Jan 03 09:57:33 * Solver encountered 1 problem(s): Jan 03 09:58:09 bitbaking only the recipe works and all files are listed in files-in-package.txt Jan 03 10:04:57 stuom1: perhaps you could say what the rootfs error actually is? Jan 03 10:05:07 stuom1: sounds like you're getting closer Jan 03 10:06:08 "Unable to install packages" then a long list of packages and the text above Jan 03 10:07:31 stuom1: there should be a clue somewhere about why its unable to install them though? Jan 03 10:08:25 there is also * Solution 1: Jan 03 10:09:06 stuom1: I'd suggest pastebinning the log somewhere so we can see the full output. You're not picking the useful lines to paste here... Jan 03 10:11:45 @RP i put it here https://pastebin.com/zc7Ttwq8 Jan 03 10:12:19 stuom1: the key line is "nothing provides lmfit needed by lumi-1.0+git0+dd5202484e-r0.aarch64" Jan 03 10:12:43 stuom1: so something in the lumi recipe is wrong as its asking for something which doesn't exist Jan 03 10:12:52 yes i thought so, thats why I pasted it :P Jan 03 10:13:10 stuom1: you didn't paste that above Jan 03 10:13:34 Interestin, it shows on my screen Jan 03 10:14:25 stuom1: something odd about lineendings and your client? Its not here :/ Jan 03 10:14:31 Maybe Jan 03 10:15:09 Bitbaking lmfit works, bitbaking lumi works, but bitbaking image give this error Jan 03 10:15:57 stuom1: right, but the recipe build probably isn't checking the dependencies fully Jan 03 10:16:10 only when you try and use them is it finding this for some reason Jan 03 10:16:39 stuom1: is this an older release? Jan 03 10:17:09 older release of what? Jan 03 10:17:47 stuom1: poky/yocto. Which release is it? Jan 03 10:21:55 poky 2.6.3 Jan 03 10:22:59 must be something wrong with my lmfit recipe then because without it everything works, but it is such a basic pypi recipe... Jan 03 10:32:57 I added setuptools3 to inherit in the recipe and now it works, interesting... I wouldn't have guessed from that error message heh Jan 03 10:36:04 i still dont understand why it was needed though and bitbaking the recipe separately worked Jan 03 10:41:35 stuom1: How are the files split up in this lmfit recipe? Does it have multiple packages? Jan 03 10:41:51 it kind of sounds like the lmfit package itself may be empty and not generated? Jan 03 10:42:36 the recipe has basically only checksums and "inherit pypi" and thats it Jan 03 10:44:04 stuom1: have a look at the generated packages in tmp/deploy/XXX and see if they have the files you expect in them Jan 03 10:46:22 * RP wonders why this sigs test keeps failing :( Jan 03 12:48:24 If I get error "aarch64-tdx-linux-gfortran: error: libgfortran.spec: No such file or directory" how can I see what exact directory it is trying to find it from? Jan 03 13:05:30 is there a way to build a single package using an older version of GCC, but keep everything else the same? Jan 03 13:06:44 s/package/recipe/ Jan 03 13:19:53 Does anybody know what has happened to libiconv? Krogoth apparently had it under meta/recipes-support but Rocko no longer has it. I am just getting some character set related warnings/errors when running samba so I am wondering if I need to fix it somehow. Jan 03 13:35:23 Hi muppe Jan 03 13:35:31 Hi Jan 03 13:35:44 I don't know about this lib Jan 03 13:35:58 but when i'm in your case i use that : https://layers.openembedded.org/layerindex/branch/master/recipes/?q=libiconv Jan 03 13:36:05 to find where the recipe is :) Jan 03 13:36:47 in your case -> meta-mingw or meta-gplv2. Hope it help ;) Jan 03 13:38:05 ok. Thanks. At least I now know where to find it. Not sure if that is really the main issue I am having. Jan 03 13:38:29 and the reason why: https://github.com/OSSystems/oe-core/commit/19b7e950346fb1dde6505c45236eba6cd9b33b4b Jan 03 13:38:35 I tried to ask about the "core" problem at #samba but I haven't gotten any replies yet. Jan 03 13:39:05 I see. Jan 03 13:40:19 But here is the main mystery I posted on Samba forum. Feel free to comment if you have a clue (or anybody else). Jan 03 13:40:20 I have two Yocto based linux distros; one running samba v4.1.12 and another one v4.6.7. I can access the shared disk on both devices using Windows 10 machine if the linux machines are connected to our local network. However, if the linux PC has a static IP address and a direct cable connection to Win10 PC, only v4.1.12 works. So I wonder what change in Samba implementation could cause this? I am us Jan 03 13:40:20 ing the same smb.conf in both cases. Maybe smthng should be changed there. Jan 03 13:41:13 @muppe no more idea sorry Jan 03 13:41:56 :( ok, no problem Jan 03 13:42:27 Any samba gurus online? Jan 03 14:07:58 Hi all, I'm stuck on qt5 error on Zeus branch (more log here: https://pastebin.com/jKgLXsu9), someone here has some info ? Jan 03 14:10:44 PinkSnake you might be running out of RAM Jan 03 14:11:05 try to add some swap Jan 03 14:11:29 hoooo nice I will tak ea look, could you please tell how you get that from the log ? I'm stupid maybe ? :( Jan 03 14:11:52 something killed your compiler Jan 03 14:11:58 i got similar error once Jan 03 14:12:14 Thanks i will try after a fresh reboot Jan 03 15:22:52 alsa didn't go away before I used MACHINE_FEATURES_remove instead of DISTRO_ or IMAGE_ Jan 03 15:23:56 RP, I saw a pod.man build failure on perl on zeus. didn't master have a similar issue? Jan 03 15:27:04 armpit: reproducible builds? Jan 03 15:30:39 not sure. let me check Jan 03 15:48:22 RP, I don't think so. its the genricx86 build Jan 03 15:48:41 * armpit ill bug it Jan 03 16:02:59 armpit: not sure then Jan 03 16:07:48 you thinking about the podfix changes Jan 03 16:11:15 armpit: right, they were repro builds specific Jan 03 18:21:37 RP: I have pushed future glibc 2.31 update here https://git.yoctoproject.org/cgit.cgi/poky-contrib/log/?h=kraj/glibc-2.31 Jan 03 18:21:51 RP: give it a shot if there is downtime Jan 03 18:22:00 on AB Jan 03 18:32:21 New news from stackoverflow: Error in building pulseaudio in yocto Jan 03 21:20:43 I'm on Krogoth so my system is using Smart package manager. Is there a good way I can tell the system in a recipe that a certain package needs to be the last one to be installed out of all the packages? RDepends the last one it usually install? Jan 03 21:22:07 The issues is b2qt-launcher was renamed from b2qt-launcher to boot2qt-launcher for which I put RREPLACES in the new recipe this causes the previous package to uninstall which unintentionally kills my app during an update Jan 03 22:08:59 not really.. packages don't have a specific install order.. the operations in them should produce an expected output, but the order itself isn't promised Jan 03 22:09:59 and krogoth, scary Jan 03 22:26:14 hmm it's an issue since b2qt-launcher will run systemctl enable b2qt.service followed by systemctl restart b2qt.service which would cause the main app to restart in the middle of an update Jan 03 22:26:28 maybe an rdepends on the app might help Jan 03 22:29:36 you might want an intermediate replacement package. that is, instead of the new name rreplacing the old one, make it just rdepend on it for now, and change the old package to no longer install anything and kill its scripts. the other package upgrading shouldn't restart the service, as postinst/prerm scripts are supposed to check for an upgrade argument to avoid restarting services on upgrade when not needed, afaik. Jan 03 22:29:38 just a random idea Jan 03 22:30:38 interesting, will try that Jan 03 22:32:33 * kergoth shrugs Jan 03 22:39:30 khem: just when I thought I might manage a green build ;-) Jan 03 22:42:29 RP: it will be green I think, I have clean world builds and all other needed patches are already upstreamed Jan 03 22:42:58 but ofcourse I am interested in runtime tests that AB does Jan 03 22:44:26 khem: right, I've just started a build with a crazy patch of my own included ;-) Jan 03 22:45:24 khem: sorry, I didn't see the musl patch v2, could you rebase that against master please? Jan 03 22:45:40 sure Jan 03 22:51:06 RP: rebased patch is on ml now Jan 03 22:57:25 i'm going to try to move from krogoth to a newer version. Is there and documentation or something I can look into for migration from using Smart to DNF? Jan 03 23:49:02 khem: thanks **** ENDING LOGGING AT Sat Jan 04 02:59:57 2020