**** BEGIN LOGGING AT Tue Sep 21 02:59:57 2021 Sep 21 11:59:00 Is there a way to get bitbake to not use inotify? Sep 21 12:09:27 disable it in the kernel :-) Sep 21 12:30:35 how would you do that? I'm having trouble finding anything about it through google, but google has become really bad at finding anything remotely obscure Sep 21 12:31:06 CONFIG_INOTIFY_USER Sep 21 12:31:18 but it's probably not a good idea to disable that Sep 21 12:31:25 lots of things use inotify Sep 21 12:31:57 what is your actual problem? Sep 21 12:52:45 that bitbake tries to use inotify Sep 21 12:53:02 literally that's it, qemu doesn't implement inotify and bitbake crashes if inotify syscalls return ENOSYS Sep 21 12:53:52 then disabling it in the kernel won't help Sep 21 12:53:57 since you'd get ENOSYS Sep 21 12:54:17 why are you running bitbake under qemu? Sep 21 12:54:25 seems a weird thing to do Sep 21 12:54:34 it doesn't really work with arm64 Sep 21 12:54:47 run it on a pc then Sep 21 12:54:52 I am Sep 21 12:55:08 so what's arm64 got to do with it? Sep 21 12:55:17 this is an arm PC Sep 21 12:55:32 run it on an x86 pc then Sep 21 12:56:00 I wanted to see if it was possible to run bitbake on this PC through emulating x86 Sep 21 12:56:04 also, arm64 pc what? Sep 21 12:56:07 I know it'd be slow and everything, it's mainly an experiment Sep 21 12:56:17 most of Apple's new PCs run ARM Sep 21 12:56:27 that's not a pc then Sep 21 12:56:29 it's a mac Sep 21 12:56:40 looks a lot like a personal computer to me Sep 21 12:58:51 I don't know what else you put in the word PC which makes a normal laptop running a unix not a PC Sep 21 13:00:26 well, some would say a pc is an x86 running windows Sep 21 13:00:42 oh, then running this on a PC would certainly not have helped, since yocto doesn't support windows afaik Sep 21 13:00:53 the more common meaning is a descendant of the ibm pc Sep 21 13:01:02 regardless of what it runs Sep 21 13:01:15 the proper solution is to _not_ run this on a PC but on a build server Sep 21 13:02:09 my 16-core epyc desktop runs it just fine, no need for a separate server Sep 21 13:02:34 yeah I'm gonna try this on my 16-core desktop too, but long-term, it would be better to have a build server for this Sep 21 13:03:35 what distinguishes a server from a desktop? Sep 21 13:03:55 mine has a "server" motherboard fwiw Sep 21 13:04:17 whether its primary purpose is to be a personal computer or to be a shared resource I suppose Sep 21 16:38:30 what's the nicest way to provide a custom bblayers.conf.sample? Is there a way to make `source poky/oe-init-build-env` automatically use my bblayers.conf.sample? Sep 21 18:27:53 Should yocto be building a custom glibc first which it uses to compile the host GNU toolchain? Currently it tries to build a GNU toolchain using whatever glibc version is already on the host system, which usually fails Sep 21 19:22:06 mort: Search for TEMPLATECONF , setting it before calling init-build first time; copy your sampels if the file is missing from your conf. Sep 21 19:33:56 FredO: yeah, I know about that, but it's a bit annoying to have to specify TEMPLATECONF every time Sep 21 19:34:05 especially since it has to be an absolute path (or relative to the poky dir or whatever) Sep 21 19:34:16 not a huge deal tho Sep 21 19:34:41 mort: i have another script using oe-init.... that sets it Sep 21 20:07:05 khem, hi Sep 21 20:07:34 I am flooded by a series of bbappend changing OE's defaults :/ Sep 21 20:08:13 glibc is maybe valid, others are very strange Sep 21 20:08:16 https://github.com/OpenPLi/openpli-oe-core/commits/hardknott Sep 21 20:08:56 this one is especially strange (undocumented) https://github.com/OpenPLi/openpli-oe-core/commit/58f533214fb1bed0b5ab71906ca8fd7b1ce28af0 Sep 21 20:13:41 FYI git.qemu.org seems to be down, I'm using a premirrors to workaround this: Sep 21 20:13:44 git://git.qemu.org/git/.* git://gitlab.com/qemu-project/BASENAME \n \ Sep 21 20:13:47 gitsm://gitqemu.org/git/.* gitsm://gitlab.com/qemu-project/BASENAME \n \ Sep 21 22:00:03 khem, and that one (your crossguid) https://github.com/OpenPLi/openpli-oe-core/pull/1056/commits/ea7b25aacd5af80b384481c3a6849db2adaf5588 Sep 21 22:00:57 unfortunately these committers never add proper messages so there is basically no history... Sep 21 22:02:42 I'll try to distill possible backports **** ENDING LOGGING AT Wed Sep 22 02:59:57 2021