**** BEGIN LOGGING AT Tue Mar 26 02:59:56 2019 Mar 26 07:38:39 o/ Mar 26 08:20:52 Morning! Mar 26 17:38:15 ok, mdev is now working with Halium Mar 26 17:59:34 proposal: https://github.com/shr-distribution/meta-smartphone/pull/95 followed by https://github.com/webOS-ports/meta-webos-ports/pull/321 Mar 26 18:00:12 ^ This has only been tested on tissot, handle with caution :) Mar 26 18:27:55 Tofe: Ah nice. This is to solve the 1 minute boot delay on Thud? And something else? Mar 26 18:30:49 just this; but it does remove the dependency of our initrd on udev, which was a bit overkill for the task... Mar 26 18:31:17 all in all, we really just need these by-partname symlinks and a devtmpfs :) Mar 26 18:32:15 At some point I might even propose a patch to the halium script to use uevent information directly Mar 26 18:41:01 Yeah uevent directly might be ideal at some point Mar 26 18:43:16 But udev just for the partitions is overkill indeed Mar 26 18:52:52 morning Mar 26 18:53:17 Tofe: the PRs are for thud just because the delay issue is seen since thud or it doesn't work with sumo for some reason? Mar 26 18:53:40 Tofe: if the first, then I would try to apply in currenly used sumo already Mar 26 18:53:59 JaMa: it's just that I debugged it on thud Mar 26 18:55:37 it should work on sumo as-is, I guess; maybe there some little merge, but nothing scary Mar 26 18:55:39 ok, will apply to thud/warrior/master and jansa/sumo to test it on sumo Mar 26 18:56:04 for the mainline hammerhead there will probably be quite a big merge there Mar 26 18:56:17 as I already reworked the initrd recipes Mar 26 18:56:50 still, it's not rocket science, the main thing is the mdev script Mar 26 19:16:52 I have small cleanup on top of your pr, sec Mar 26 19:16:56 and 1 question Mar 26 19:18:53 sure Mar 26 19:19:09 https://github.com/shr-distribution/meta-smartphone/pull/95/commits/9ad15905590f2867b322c8077bab0f84d390db5b#diff-8b458bc4d78e7c1a6fde73038776b86dR6 Mar 26 19:19:28 I've removed the trailing whitespace and unified indentation to 4 spaces Mar 26 19:19:40 but the ACTION=add looks a bit strange Mar 26 19:22:24 https://github.com/openembedded/openembedded-core/blob/master/meta/recipes-core/busybox/files/mdev-mount.sh#L11 if you see a use there.... Mar 26 19:22:35 I think we can safely remove that line. Mar 26 19:23:26 I don't know the internals of mdev, so I tried to stick to the template of the script Mar 26 19:23:52 unless these .sh files are combined in something longer I don't see any Mar 26 19:24:18 Let me re-push without it, if it's not too late Mar 26 19:24:24 too late Mar 26 19:24:27 ok :) Mar 26 19:24:29 made the change locally :) Mar 26 19:38:54 Tofe: meta-webos-ports merged with similar cosmetics on top, lets hope I didn't break anything, starting the build now Mar 26 19:43:55 :) ok thanks ! Mar 26 19:46:02 thanks to you!; any other blocker for thud upgrade? Mar 26 19:46:29 or just lack of testing on other MACHINEs? Mar 26 19:46:36 mainly yes Mar 26 19:51:33 btw: I've stopped rebasing the meta-webos-ports branches and doing the cherry-picks like in meta-smartphone Mar 26 19:52:05 the obvious advantage is that they are no longer force pushed (and don't include changes from jansa/* branches) Mar 26 19:52:58 but disadvantage is that if someone else is doing the merge through github PR it's easy for me to forget to cherry-pick to newer branches (and/or more difficult for the person merging the PR) Mar 26 19:53:36 but I'm feeling optimistic that we'll soon upgrade to warrior and there will be only 1 active branch + master Mar 26 20:09:28 but warrior is still a moving target, isn't it? Mar 26 20:13:19 it moves more than sumo or thud, but it's already in feature freeze stage, so no major component upgrades are expected Mar 26 20:13:42 sumo and thud were updated 3-4 times in last 2 days, but that's a bit exception Mar 26 20:14:05 Yocto will do another point release and it should get a bit more quiet again Mar 26 20:14:19 e.g. the busybox fix for the unsecure symlinks was merged today Mar 26 20:19:48 ok I see; then yes warrior can be an interesting target Mar 26 20:20:22 though I wanted to do an intermediate LuneOS release in-between, on sumo (or thud), with devices confirmed working well Mar 26 20:22:05 there should be new warrior build ready on jenkins tomorrow (tissot, qemux86, hammerhead) if it works the same as on thud then it would be worth it Mar 26 20:22:23 well there is still the gobject-introspection issue on hammerhead/warrior Mar 26 20:22:47 but I think we don't really use gobject-introspection anywhere, so I can easily just disable it Mar 26 20:23:04 let me do that for tonights build at least so we get image Mar 26 20:23:11 ok :) Mar 26 20:25:49 well that won't help because some other postinst will fail to run inside qemu-arm during do_rootfs and cause image build to fail, but at least we should get a bit closer and I know what to debug tonight :) Mar 26 23:39:39 heh, the qemu-arm segfault is caused by ld-2.19 inside recipe-sysroot **** ENDING LOGGING AT Wed Mar 27 02:59:57 2019