**** BEGIN LOGGING AT Sat May 16 02:59:57 2020 May 16 05:01:37 Morning May 16 05:13:21 Tofe: It shouldn't be really. I forked, did my changes and PR'ed and bumped to SRCREV after my PR was merged, not even to latest May 16 06:58:01 Tofe: It could be somehow related to this commit, but that has been there fore a while already: https://github.com/webOS-ports/qtsensors-sensorfw-plugin/commit/e9b28f3d6ad64f1c2bc88adaf59781cfd45e9455 May 16 07:43:09 Tofe: If I revert that commit it builds... May 16 08:08:36 Herrie: ah, well, let's just revert it then ? I don't really see how using QMAKE_USE is such an improvement anyway May 16 08:09:16 it relies on .json files being provided by sensorfw May 16 08:13:10 Tofe: I think I tested back with 5.12 seems something might have broken with 5.14 upgrade somehow? May 16 08:13:49 that could be May 16 08:15:24 Or something on sensorfw side but don't really see what May 16 08:45:52 https://git.sailfishos.org/mer-core/sensorfw/commit/2926a491dd603a7ad482a928440efba63caa2acb I hope this won't be a problem May 16 08:51:54 Herrie: it looks like we need, to have QMAKE_USE work, to define things like QMAKE_LIBS_sensorfw . It also looks like it's not documented at all. I'd prefer stick to the pkgconfig variant. May 16 09:00:16 Tofe: I'll double check the 5.14 plugin sample code, maybe they changed something May 16 09:00:27 Herrie: I think I understand what is wrong with the QMAKE_USE here: the imported commit is only partial, and we'd also need to patch a configure.json file to have it working https://github.com/qt/qtsensors/commit/5bf1bd00335d0c95c54de426a02da2a98257fedb May 16 09:00:40 Weird thing is we're not having this on builder it seems May 16 09:01:08 so better just not import that commit at all May 16 09:01:13 I took Zeus from meta-webos-ports and reapplied all my ACG stuff, not sure how/why this now breaks May 16 09:01:39 Anything in rpm on sfos side is their packaging May 16 09:01:50 So nothing to worry about for now May 16 09:02:20 ah, the commit comes from Mer ? May 16 09:08:42 -- I guess we did the split to have a generic qtsensor build, and a MACHINE_ARCH plugin part ? I think I asked that last year too May 16 09:09:51 yes, back in september :) May 16 09:10:32 "Sep 24 19:51:29 ah, luna-sysmgr would become MACHINE_ARCH... that's not nice" May 16 09:11:21 I'll put that in our README May 16 09:22:55 Tofe: ehm well the plugin we take 1:1 from Qt I think from their sample code. Instead of forking we could possibly just point to the folder in our recipe will have a look sometime May 16 09:36:43 Herrie: that still wouldn't build with the aforesaid commit; we'd need to build the whole thing and just handpick the plugin binaries May 16 09:39:25 Tofe: You mean when you add the missing bits from Qt commit (Json etc) it still doesn't build? May 16 09:41:11 No, I mean that you can't just build the sub-folder of the plugin, if you want the .json to be taken into account it has to be the whole qtsensors build May 16 09:42:22 though it would be worth a try, I didn't really try :) May 16 09:46:55 However, rebuilding the whole qtsensors source code isn't such a big thing, if we want to go that way May 16 09:59:09 For the moment, anyway, we should just revert the little commit, whatever we choose next May 16 10:56:33 Herrie: in first use, I don't get any language list May 16 10:57:02 Tofe: Herrie: do you want me to merge Qt 5.15 in zeus (and forget about 5.14)? May 16 10:57:39 Herrie: "user.warning LunaAppManager 2360 (LunaAppManager) PBNJSON src/pbnjson_c/jschema.c:250 :: Unable to open schema file /etc/palm/schemas/localization.schema" seems to match the problem May 16 10:59:48 Tofe: I think that warning existed before and is not the cause May 16 10:59:51 mmmh I get lots of pbnjson issues like "user.crit rdxd 1434 (rdxd) PBNJSON src/pbnjson_c/jobject.c:674 :: Attempt to cast null 0x7f9bff93f0 to object" May 16 11:00:06 Not 100% sure though May 16 11:00:09 not necessaily related, still looking at logs May 16 11:01:34 LunaWebAppManag[3148]: LocalePreferences::getLocaleInfoCallback(): [{"returnValue":false,"errorCode":-1,"errorText":"Denied method call \"getPreferences\" for category \"/\""}] could be related, but firstuse is a QML app though May 16 11:02:52 Yes that should be related May 16 11:04:38 LunaSysMgr[2356]: LocalePreferences::getLocaleInfoCallback(): [{"returnValue":false,"errorCode":-1,"errorText":"Denied method call \"getPreferences\" for category \"/\""}] this one seems spot on May 16 11:05:03 let's continue the test anyway May 16 11:06:10 I solved 100's of those permission errors May 16 11:06:21 :) I guess yes May 16 11:06:26 I recall there were a handful that need sorting still but got stuck May 16 11:07:27 I should have them written down somewhere May 16 11:07:28 it also fails to activate wifi, but it well be the same thing May 16 11:07:38 +can May 16 11:07:38 But they're easy enough to spot in logs May 16 11:08:05 Lunasysmgr from owo didn't have localization so that json is false lead May 16 11:08:20 If you check lunasysmgr recipe you see it's commented out there May 16 11:08:22 you're right, I don't see it either in my old pinephone image May 16 11:08:43 no background image, but that's stored in the preferences. May 16 11:09:38 sound ok, switch to normal mode ok, rotation sensor ok, ofono ok (no visible reg, at least) May 16 11:12:06 starting the messaging app isn't working, not much in the logs May 16 11:12:45 let's reboot once May 16 11:12:49 Tofe: Sounds like what I had as wlel May 16 11:12:57 I mean it gets far, there's a few minor things to fix soon May 16 11:13:01 Mainly the preferences stuff May 16 11:13:02 It seems May 16 11:13:56 the preferences is the most visible one yes; I don't know yet why web apps don't start, but it's probably one obvious thing May 16 12:07:07 Tofe: I used qemu for all this in general. Easy to scp the json files over, bit more reliable vs adb imho May 16 12:07:30 for the webapps it seems that the qtwebengine processes aren't started at all May 16 12:09:29 could it be just because of getpreferences too ?... why not **** ENDING LOGGING AT Sun May 17 02:59:57 2020