**** BEGIN LOGGING AT Sun Feb 04 03:00:02 2018 Feb 04 09:40:14 Morning Feb 04 09:40:23 Tofe: Seems my persist on mido doesn't get mounted Feb 04 09:41:37 Morning Feb 04 09:42:03 Herrie: what's the difference between your custom android-gadget-setup and the generic one ? Feb 04 09:42:47 Tofe: Well I just addded your patched bits Feb 04 09:42:51 Which somehow don't seem to stick Feb 04 09:42:56 Just for quick debugging I now added it Feb 04 09:43:06 For luna-next, frankly, right now, I'm not sure how to proceed; there's no error message, nothing. The only little clue is that the qpa hwcomposer returns a wrong physical screen size, but I doubt it's significant Feb 04 09:43:27 Tofe: Well I have rild and pulseaudio acting up as well Feb 04 09:43:56 Seems that the persist partition doesn't get mounted Feb 04 09:44:20 Seeem we have barrier=1 in fstab for base-files however I read that " Note: Logical Volume Management (LVM) based partitions cannot use the barrier option. " Feb 04 09:44:25 So I guess I should remove that :P Feb 04 09:57:27 ah, yes, good catch Feb 04 10:16:56 oh wait, on TP, there is a "luna-next[1670]: library "libadreno_utils.so" not found" Feb 04 10:22:58 Tofe: I guess we need that one :P Feb 04 10:23:23 yes, but that shouldn't have changed with newer Qt... Feb 04 10:37:22 true Feb 04 10:41:38 ok so it's not the physical screen size Feb 04 10:42:19 Tofe: OK seems I'm trying to mount 016 instead of 026 :P Feb 04 10:42:21 That doesn't help Feb 04 10:42:50 ahah, yes :) Feb 04 10:58:15 damn, build.webos-ports.org is slow today Feb 04 12:41:27 How long does a clean build of a LuneOS image take for you at the moment? (General question.) Mine last night took over 13 hours. I wonder how I might speed that up. Feb 04 12:45:36 6-7 hours Feb 04 12:46:03 For me SSD was the key. Also, I build it in a chroot, not in a VM Feb 04 12:49:24 Tofe: How many cores do you have? Feb 04 13:06:04 4 Feb 04 13:06:08 an i5-6600 Feb 04 13:07:18 elvispre: I got an AMD X6-1055 with 8GB ram and 500 GB SSD Feb 04 13:07:22 It's not a lot quicker Feb 04 13:07:31 Both qtbase & qtwebengine take long Feb 04 13:10:00 Adn that's on native Ubuntu 16.04 Feb 04 13:10:08 Tofe: test_vibrator works on mido Feb 04 13:10:13 Most others fail ahrd though :P Feb 04 13:10:49 Herrie: well, that's a beginning :) Feb 04 13:11:08 Herrie: can I see systemctl status output? Feb 04 13:11:11 But I remember the tests were flaky anyway, werne't they? Feb 04 13:11:59 bshah: https://bpaste.net/show/374f0e02aff2 Feb 04 13:12:12 I get no image, though most other things in log seem OK Feb 04 13:12:17 not what I asked :P Feb 04 13:14:37 Just rild fails and pulseaudio as main Feb 04 13:14:45 bshah the regular segmentation faults in general Feb 04 13:15:31 Tofe: Herrie: I have a 4-core i5-760. Maybe getting a bit long in the tooth. Stacks of RAM and huge spinny disks. Feb 04 13:18:09 bshah: Any test output you're looking for specifically? Feb 04 13:18:29 Herrie: I think he just would like to have "systemctl status" output Feb 04 13:18:29 I am looking for systemctl status to see if your container is happy or not Feb 04 13:18:34 :P Feb 04 13:19:13 elvispre: the build does quite a lot of tiny disk access, and many files are used only once. In this use case, SSD do a much better job than any spinning disk, I think. Feb 04 13:21:20 Tofe: Interesting. Thanks. Feb 04 13:22:11 Herrie: for TP I'm investigating a possible cause within the hwcomposer. It could eventually be triggered by some slightly different Qt code (like, creating a temporary window of size zero). It's not really the kind of code anyone is happy to debug, though... Feb 04 13:25:02 (systemctl status tells me my phone has been up for 45 years! :-) Feb 04 13:29:25 Tofe: Yeah those qcom bits aren't fun Feb 04 15:59:37 JaMa: ping Feb 04 16:06:21 JaMa: nevermind, I found the solution Feb 04 16:30:16 JaMa: https://github.com/shr-distribution/meta-smartphone/pull/54 Feb 04 16:30:32 Herrie: ^ this fixes our display issue Feb 04 16:30:54 The Qt upgrade was wrong lead, in the end :) Feb 04 17:59:46 Tofe: That must have been me then :P I guess I might try it on my mido, might help there too :P Feb 04 18:00:01 Sorry for that didn't think of it Feb 04 18:08:57 Herrie|Pre3: I didn't think of it either :p And it was me who said we should put a preferred version, back in the days when I merged the halium headers :p Feb 04 18:09:34 Anyway, bluetooth stll doesn't initialize, unfortunately Feb 04 18:09:40 (bluez5 I mean) Feb 04 18:12:32 Tofe: LOL, well it might help for mido & onyx, who knows, we'll see Feb 04 18:13:06 Herrie|Pre3: I think bb chooses the highest version number Feb 04 18:13:16 so it shouldn't change anything for these Feb 04 18:13:31 and for mako & hammerhead... we were just lucky Feb 04 18:13:44 we don't use that much of the API, so we didn't even see the change Feb 04 18:14:17 And tenderloin is the only machine where we define QCOM_BSP, which actually changes the ABI Feb 04 18:18:24 OK Feb 04 18:18:39 Well it doesn't hurt to try Feb 04 18:22:35 Sure Feb 04 18:24:45 You know, I'm not sure I'll be able to bring bluez5 to the TP... It's quite a new area to me, and I'm not even sure that the backport that was done for N4&N5 will work for TP Feb 04 19:32:57 Tofe: Since most of this backporting for N4/N5 was done by morphis, might be worth to ping him, he might have some thought/pointers **** ENDING LOGGING AT Mon Feb 05 03:00:02 2018