**** BEGIN LOGGING AT Sat Jun 09 03:00:07 2018 Jun 09 05:16:35 Morning Jun 09 05:18:43 What / who builds the content of http://build.webos-ports.org/? Jun 09 05:23:16 I see jenkins jobs on http://jenkins.nas-admin.org/ for the luneos-* folders (I assume) but nothing for the halium-luneos-* ones. Jun 09 05:25:11 Are the build scripts for halium-luneos configured somewhere public? Jun 09 06:32:05 elvispre: it's the halium-luneos-* jobs that take care of that Jun 09 06:37:50 elvispre: I think the scripts are not in public, but I pasted them before Jun 09 06:37:59 Can do that again Jun 09 06:41:09 elvispre: https://bpaste.net/show/35c61bac62c7 Jun 09 06:41:45 elvispre|s: I use something similar locally for building halium bits Jun 09 06:42:06 Just we'd need to tweak it a bit still Jun 09 06:42:25 Since we replicate a lot of the halium android manifests. There must be a nicer way Jun 09 06:44:31 Herrie: I have a copy of that script locally (you've shown me it a couple of times before). I just wondered if it was current as you were talking about filtering out content with regexes and so on. Jun 09 06:45:46 elvispre|s: Not yet really Jun 09 06:46:11 We should I guess update it, I'm still in the middle of that but got sidetracked with OSE and other stuff Jun 09 06:46:21 Ideally we get rid of https://github.com/webOS-ports/android/blob/luneos-halium-7.1/default.xml completely Jun 09 06:47:14 And just keep https://github.com/webOS-ports/android/blob/luneos-halium-7.1/luneos_targets.xml where we remove the hybris-boot as per https://github.com/webOS-ports/android/blob/luneos-halium-7.1/luneos_targets.xml#L3 Jun 09 06:47:28 The whole devices we already migrated to halium-devices Jun 09 06:47:43 So we don't need our device bits anymore in luneos_targets Jun 09 06:48:30 What I do currently to build locally is to follow https://docs.halium.org/en/latest/porting/get-sources.html#initialize-and-download-source-tree Jun 09 06:48:43 And then to use a slightly modified script locally Jun 09 06:48:49 Tofe: Ah! I missed those jobs. I was looking in the jenkins project group tabs and managed to ignore "All" :-) Jun 09 06:51:13 Herrie: Good information, thank you. (I was more or less correct then!) Jun 09 06:51:29 Yes you were Jun 09 06:51:38 I need to find some time to test it properly and do a PR Jun 09 06:51:51 Just now we need to update our full manifest every time halium updates theirs Jun 09 06:51:57 Doesn't make a whole lot of sense Jun 09 07:00:31 Herrie: best would be to have a sort of "include" of the one form Halium Jun 09 07:00:34 from* Jun 09 07:20:12 morning Jun 09 07:36:45 saidinesh5: it's not really an API issue; it's just integration the PalmServiceBridge IDL into chromium Jun 09 07:37:12 But the more I look at it, the more I think we could handle it externally Jun 09 07:37:28 And that would be a great relief :p Jun 09 07:45:04 Tofe: Yeah, well we could take their Android repo and then in luneos.xml just remove the bits we don't need basically Jun 09 07:45:13 Similar like we do in luneos_targets.xml currently Jun 09 07:45:23 Which is mainly hybris-boot/halium-boot Jun 09 07:47:16 Just need to make sure I upstream the missing devices we have in luneos_targets.xml to halium-devices Jun 09 07:47:26 I'll have a look at this when I find some time in next couple of days Jun 09 07:47:38 Herrie: ok :) Jun 09 07:47:51 Note that there actually is an include statement in the manifest syntax Jun 09 07:48:44 so if we checkout their manifest, and then ours, the latter could begin by including theirs (from a subfolder for instance) Jun 09 07:49:26 pretty sure we can end up with something simple that way Jun 09 07:55:05 Tofe: Ah that's good to know! Jun 09 07:55:09 Will look into that! Jun 09 16:33:56 anyone ran a newer systemd with these old kernels? Jun 09 16:34:52 seen something about.. "Failed to determine whether /sys is a mount point: Bad file descriptor" then it freezes the API filesystem or something Jun 09 16:36:17 OrokuSaki: can you define "newer" ? and "old" ? Jun 09 16:37:32 Sup dude.. long time! Yeah 3.4 kernel with say systemd 234... I am trying 233-6 at the moment.. running archlinux.. I had a 2 year old image that I like to mess with libhybris on.. and went to upgrade it and ran into that error.. so I am trying 233 as of now.. waiting for it to finish Jun 09 16:37:53 I googled this issue and someone ran into it awhile back http://logs.nslu2-linux.org/livelogs/webos-ports/webos-ports.20171006.txt Jun 09 16:38:29 so far we are on 232 Jun 09 16:39:13 nod.. I think it has something to do with MountAPIVFS Jun 09 16:39:28 Did you try https://github.com/webOS-ports/meta-webos-ports/blob/master/meta-luneos/recipes-core/systemd/systemd/fd_fdinfo_mnt_id_disablefdinfostat.patch ? Jun 09 16:39:43 I think this is a systemd bug.. no I will try patch though.. thanks Jun 09 16:40:34 It's not really a bug, but systemd expects the kernel to have some working functionalities, which appear in later kernels Jun 09 16:41:07 In our case, fdinfo was there, but only partially working Jun 09 16:41:10 Yeah... but the requirements for systemd say you need at least a 3.4 kernel.. Jun 09 16:41:27 oh, no, they say you need a much newer one :) Jun 09 16:41:35 I don't think they actually test 3.4 kernels though.. =) Jun 09 16:42:05 https://github.com/systemd/systemd/blob/v232/README#L38 Jun 09 16:42:26 bumped to 3.13 for v234 Jun 09 16:43:56 Though it's not a hard requirement: they just say 5+ years kernels are not officially supported. Jun 09 16:44:49 Ahh Jun 09 16:45:07 I like the 3.13 kernel more then 3.4 for the touchpad Jun 09 16:45:17 Or it is 3.10... Jun 09 16:45:33 3.10 I think; but I never tested it. Did you? Jun 09 16:45:56 yeah but only on sailfish which used a really old systemd Jun 09 16:46:51 I am curious how xwayland works now.. I had a bug before where the touchscreen worked one time with this archlinux image Jun 09 16:47:18 Do you still have the link to the sources of the kernel they used ? A newer kernel for touchpad would be quite great for the future Jun 09 16:48:15 Well.. I was the porter for the sailfish stuff.. Think it was just a flintman kernel... Only thing was they had a kernel configuration utility that would automatically change the kernel config for ya Jun 09 16:49:14 I ran into a bug with Qt5 and orientation with certain qt5 programs... I think it is because the touchpad's kernel compass may be jacked Jun 09 16:50:18 ah ok yes Jun 09 16:50:24 I remember flintman's kernel Jun 09 16:51:09 Nod... I have a archlinux image that does libhybris and freedreno at the same time.. Really weird... Wanted to compare glxgears Jun 09 16:51:12 =) Jun 09 17:34:32 "0: qtwebengine-5.11.0+gitAUTOINC+28e0320235_1785e2c1eb-r0 do_compile (pid 3995) 95% " ... suspens... Jun 09 18:13:56 JaMa: looks like we have a mismatch between USE_OZONE in QtWebEngineCore and use_ozone in chromium Jun 09 18:14:17 It's causing a link error here, the GetFontFamilies_SlowBlocking symbol is defined twice Jun 09 18:14:37 Apart from that it seems to build Jun 09 18:19:24 JaMa: here's what I used: https://github.com/Tofee/meta-webos-ports/commit/3f89630b93ee0009bc5110fd30710275471bbb85 on top of you jansa/pyro branch Jun 09 18:19:31 that'll be all for tonight !* **** ENDING LOGGING AT Sun Jun 10 03:00:04 2018