**** BEGIN LOGGING AT Thu Nov 01 03:00:00 2018 Nov 01 05:58:32 Morning Nov 01 07:42:49 Morning! Nov 01 07:43:36 elvispre: if you flash back android, make sure /data/time/ats* is deleted, does it regenerate it ? Nov 01 09:03:40 Ah LG updated their Blog post with the October release: http://webosose.org/blog/about-webos-ose-releases/5th-release-2018-october-2018/ Nov 01 09:20:12 Seems ZRAM support is there Nov 01 09:20:24 That might be useful for targets like TP which have little memory Nov 01 09:32:58 yup Nov 01 10:19:55 Herrie|Laptop: I found out that in the ls2 roles we are missing all the permissions for the empty ("") service name Nov 01 10:20:09 why we try to register such a name is a wonder in the first place, but well Nov 01 10:20:34 Tofe: There must have been a reason for this.... Nov 01 10:20:43 Which probably is only known to morphis :P Nov 01 10:21:11 one reason I found is that we have a LSRegister in luna-syscommon, where it doesn't know what app it's dealing with Nov 01 10:22:49 Here's the list of impacted files: https://bpaste.net/show/9bc95c620996 Nov 01 10:24:08 So when I define an ad-hoc permission (all permitted), there are much less ls2 warnings in the log, and it becomes much clearer Nov 01 10:34:57 Morning! Nov 01 10:36:05 Tofe: I woke up (just now!) planning to do just that. Nov 01 10:36:30 (I am on holiday for a few days.) Nov 01 10:42:52 Tofe: Ah clear, that makes sense Nov 01 10:43:07 When luna-sys-common doesn't know the app, you need to have something generic Nov 01 10:43:35 yes, well, I still would have preferred a non-empty name Nov 01 10:43:56 But it'll do for now Nov 01 10:50:37 Tofe: You're now looking into OSE already or.... ? Nov 01 10:50:52 yep! Nov 01 10:51:18 I have some time this week-end, so I take the occasion Nov 01 10:58:02 I wonder if adding a permission with an empty service name if backward compatible Nov 01 10:58:08 Should be, I guess Nov 01 11:09:39 Tofe: It's probably some hack that existed previously or was added by morphis somewhere Nov 01 11:24:45 Tofe/elvispre: Sent you email with release announcement draft Nov 01 11:24:57 thanks! Nov 01 13:15:27 Herrie|Laptop: Tofe: will do the PR soon Nov 01 13:15:38 JaMa: Thnx! Nov 01 13:15:53 where did you see the Qt 5.9 comment? I thought it was one recent commit I did, but that one isn't public yet :) Nov 01 13:19:38 JaMa: "Oct 31 10:06:07 Herrie|Laptop: oooh do you see whay I see ? https://github.com/webosose/qml-webos-bridge/commit/71a481132a9235e6c79b71b74dd7be1daebc6bcf#diff-3558f2d51a0d4859e7d6252b116fea31R150" Nov 01 13:53:28 ah, much deeper than my meta-webosose commit :) Nov 01 16:08:41 JaMa: maybe you can help me with an easy thing Nov 01 16:09:21 There's a qtlocation luneos plugin, currently inserted into qtlocation with a rough patch, that I'd like to build simply as a plugin in a separate recipe Nov 01 16:09:45 https://github.com/Tofee/qtlocation-luneos-plugin <-- here's my very first try, which doesn't build of course :) Nov 01 16:10:21 qmake complains with "Could not find qmake spec 'linux-oe-g++' "... I must have forgotten something basic ? Nov 01 16:10:52 Here's the recipe I'm using https://bpaste.net/show/8031f16a4e48 Nov 01 16:11:32 oooh maybe the depends... Nov 01 16:12:19 yup. ok, nevermind. Nov 01 16:25:09 :) Explaining a problem to someone (even if they're not there to nod their head) often does the trick. Nov 01 16:27:13 yes, now it builds fine, I'll be able to put away the dependency between luna-service2 and qtwebengine :p Nov 01 16:28:01 Herrie: could you please mirror/fork/whatever https://github.com/Tofee/qtlocation-luneos-plugin to our webos-ports projet on GH ? Nov 01 16:28:29 (or can I do that myself ?...) Nov 01 16:28:42 nope Nov 01 16:33:06 Tofe: Forked it, let me see if I can delete reference to your upstream Nov 01 16:38:22 Ah I should import instead of fork it seems as per https://stackoverflow.com/a/47974224 Nov 01 16:39:02 ah maybe, I never did such thing myself Nov 01 16:39:05 Tofe: You should have access now :) Nov 01 16:39:11 To: https://github.com/webOS-ports/qtlocation-luneos-plugin Nov 01 16:39:16 thanks ! Nov 01 16:41:30 Herrie: I'm trying out a build where the LS2 security isn't disabled Nov 01 16:41:57 Tofe: OK :) Nov 01 16:42:04 (and that was trigger a qtwebengine rebuild because of the qtlocation patch, so that's why) Nov 01 16:42:10 +ing Nov 01 16:42:25 Yeah that was an extremely nasty hack to get rid of some of the LS2 most obvious "blockers" Nov 01 16:42:30 Not a proper solution at all Nov 01 16:42:53 Ah so now we don't have to rebuild qtwebengine anymore in case of qtlocation patch? Nov 01 16:42:59 Well, I'm not sure if these issues are gone now or if it just doesn't print it, but I don't get that many warnings now Nov 01 16:43:13 That helps at least Nov 01 16:43:37 I've made the patch an independant plugin, so that lifts the dependency of qtlocation on luna-service2 Nov 01 16:43:52 And I've introduced a new recipe Nov 01 16:44:10 (and dropped the bbappend on qtlocation) Nov 01 16:44:43 we'll see when my build finishes, in an hour or so... Nov 01 16:45:11 I can always prepare the PR for the qtlocation plugin Nov 01 16:46:39 True Nov 01 16:56:47 done Nov 01 16:57:12 It's not a must-have for the release though. Nov 01 16:58:31 Tofe: This was tested to work OK? Nov 01 16:59:26 nope, not yet Nov 01 16:59:34 it just builds Nov 01 17:34:35 hm, webosose.org is now blocked in Russia... ?? Nov 01 17:37:00 has someone here been poking at my service simulation? i notice the traffic has gone up significantly this week, curious why. 92 clones ! Nov 01 18:12:23 EricBlade: not me Nov 01 18:12:51 nizovn: It's slow here too Nov 01 18:12:56 So could be a server issue Nov 01 18:13:03 Or really blocked in Russia, not sure :P Nov 01 18:14:41 Tofe: do you have dependency on qtbase? Nov 01 18:15:27 I don't see it in the recipe and qmake5 adds only qtbase-native iirc Nov 01 18:18:19 hmm and now I read the rest of backlog to see that I'm few hours too late :) Nov 01 18:18:28 Herrie: i get redirect page that states it's blocked. (if anyone intereseted: http://89.185.75.227/451/1f50893f80d6830d62765ffad7721742.html ) Nov 01 18:20:30 I got some ugly stretched banner, quick enough Nov 01 18:20:39 nizovn: I cannot read Cyrillic but I get the message! Nov 01 18:21:12 looks exactly the same as p0rn blocker in Korea :) Nov 01 18:21:13 :) Nov 01 19:20:03 Looking at RTC and persisting the system time offset on tissot. Nov 01 19:21:41 If I reflash the stock ROM and boot it, /persist/time/ats_1 and ats_2 both get updated. ats_12, 13, 15 and 16 do not. (I wonder where they come from.) Nov 01 19:23:08 To find that out, I have to reflash the stock ROM, NOT boot it, fastboot boot TWRP and use that to sideload LuneOS. Phew! Nov 01 19:27:11 A nice side effect of doing this is that, suddenly, the on-screen keyboard has started to work for me again. Nov 01 19:42:39 LuneOS (nor its Halium / android subsystem) does not appear to write anything at all to the /persist partition. Nov 01 19:52:35 Herrie: without the security bypass, it's slightly worse, but not that bad Nov 01 20:24:54 To give an idea of the amount of work needed, here are all the ls-hubd errors on startup: https://bpaste.net/show/8c92d919e585 Nov 01 20:26:06 The only one that I find puzzling is : [pmlog] ls-hubd LSHUB_ROLE_FILE {} Can not find service "com.palm.keymanager" permissions for executable "/usr/bin/node" Nov 01 20:26:38 (^node should be a particular case...) Nov 01 21:30:33 Tofe: Yeah there are quite some that need fixing :S Nov 01 21:30:46 That looks similar to what I had based on memory Nov 01 21:33:21 well if one looks carefully, there isn't *that* much to do: tweaks.prefs, qml-launcher, phone app... and migrating all the deprecated roles Nov 01 21:39:10 JaMa: does https://www.phoronix.com/scan.php?page=news_item&px=Mesa-EGL-Device-Extensions change anything regarding our Mesa upgrade issue? **** BEGIN LOGGING AT Thu Nov 01 23:10:38 2018 **** ENDING LOGGING AT Fri Nov 02 02:59:58 2018