**** BEGIN LOGGING AT Wed Aug 30 03:00:00 2017
Aug 30 05:16:23 <Herrie|2> Morning
Aug 30 05:16:38 <Herrie|2> Tofe: You get these too sometimes? https://bpaste.net/show/5dcd6b4e6ae8
Aug 30 05:23:47 <Herrie|2> Tofe: You know if we use crash-handler still?
Aug 30 05:23:58 <Herrie|2> https://github.com/webOS-ports/crash-handler
Aug 30 05:24:11 <Herrie|2> It gives me some warning during build about GNU HASH in ELF binary
Aug 30 05:24:27 <Herrie|2> But seems it still has upstart info in it, so looks like we're not using it
Aug 30 05:24:53 <Herrie|2> Not sure if we'd want to use it or not anyway? http://elinux.org/Crash_handler
Aug 30 05:30:18 <Herrie|2> Tofe: These errors are bpasted above are gone when I run bb -f componentname
Aug 30 05:30:27 <Herrie|2> So not a big issue
Aug 30 05:30:48 <Herrie|2> Just that I had 20 or so of them when I build 1 ARM target after another... I.e. Hammerhead after Tenderloin etc
Aug 30 05:31:51 <Herrie|2> Seems QT related, since it's all QML etc stuff (luna-next, voicecall, nemo-qml-plugin-dbus, browser etc)
Aug 30 06:44:29 <Tofe> Morning
Aug 30 06:45:18 <Tofe> Herrie|2: no, I've never seen that one
Aug 30 06:45:41 <Tofe> for crash-handler, I don't know
Aug 30 06:46:13 <Herrie|Pre3> Tofe: OK. I'll ask JaMa he might have a clue since he often does multiple machine builds.
Aug 30 06:46:32 <Herrie|Pre3> Can "solve" it with rm -rf tmp-glibc but that's nasty.
Aug 30 06:47:18 <Herrie|Pre3> I only see a reference to crash-handler in packagegroup, nowhere else, which leads me to believe it's built and deployed but not used.
Aug 30 06:47:47 <Herrie|Pre3> Seeing it would require a systemd config if previously it was using upstart.
Aug 30 06:48:02 <Herrie|Pre3> It would need a small kernel patch as well to run.
Aug 30 06:48:13 <Herrie|Pre3> Not sure we have that one I'll check.
Aug 30 06:48:28 <Herrie|Pre3> If you say it's not useful we could drop it altogether
Aug 30 06:50:56 <Herrie|Pre3> Tofe: Doing a fresh install of QT here. You run it on Windows machine too if I recall correctly?
Aug 30 06:51:21 <Herrie|Pre3> Just wondering which compiler/kit to use... MSVC I guess would be best?
Aug 30 06:51:30 <Herrie|Pre3> Or MinGW would work too?
Aug 30 07:17:05 <Tofe> Herrie|2: I'm using VS2015 currently
Aug 30 07:17:24 <Tofe> iirc, it's mandatory in order to get QtWebEngine
Aug 30 07:17:35 <Tofe> (or was it VS2013 ?)
Aug 30 07:22:12 <Herrie|Pre3> Tofe: OK I'll do VS2015 too then
Aug 30 10:56:28 <Herrie|Pre3> JaMa: morning!
Aug 30 11:08:53 <Tofe> JaMa: Morning!
Aug 30 11:09:08 <Tofe> (I must say, it's a pretty interested "Morning" I'm doing here ;) )
Aug 30 11:16:36 <Tofe> JaMa: I'm wanting to activate the full debug symbols for qtwebengine, but I'm struggling a bit...
Aug 30 11:17:02 <Tofe> So far I've added the following in our bbappend: EXTRA_QMAKEVARS_PRE += " CONFIG+=webcore_debug CONFIG+=force_debug_info "
Aug 30 11:17:27 <Tofe> it seems to do what I want, but I get build errors because the generated Makefile is incorrect
Aug 30 11:18:57 <Tofe> things like "-O2\ -pipe\ -g\ -feliminate-unused-debug-types\ -fdebug-prefix-map=....." appear
Aug 30 11:19:21 <Tofe> I trying to track it down, but without much success so far
Aug 30 11:24:34 <Tofe> Maybe you know of some other way, or have already seen this kind of issue?
Aug 30 11:25:22 <JaMa> Herrie|Pre3: Tofe: Morning
Aug 30 11:25:50 <JaMa> Tofe: doesn't ring a bell, is it with 5.9?
Aug 30 11:26:57 <JaMa> Herrie|Pre3: I've tried booting the vmdk image with qemu and if I pass the -kernel parameter directly to /boot/vmlinuz copied out from vmdk image I get the image to boot (with only the module loading service failing due to missing vbox devices as expected in QEmu)
Aug 30 11:27:19 <JaMa> Herrie|Pre3: that doesn't explain the sshd issue, but maybe it doesn't boot at all in VBox now
Aug 30 11:27:36 <Tofe> JaMa: yes, 5.9
Aug 30 11:27:44 <JaMa> my test was with Rocko which is using wic to create vmdk, so it might be different issue completely
Aug 30 11:27:46 <Tofe> JaMa: with 5.8 I get all the symbols alright
Aug 30 11:29:09 <Tofe> JaMa: this https://github.com/qt/qtwebengine/blob/5.9/src/core/config/common.pri#L58 is what I identified as the cause
Aug 30 11:30:16 <JaMa> do we build qtwebengine as release or debug?
Aug 30 11:30:28 <Tofe> I... don't know :) let me see
Aug 30 11:30:29 <JaMa> at least in chromium it should be enough to change this release flag
Aug 30 11:30:50 <Tofe> well, there's also this: https://github.com/qt/qtwebengine/blob/5.9/src/core/config/common.pri#L54
Aug 30 11:31:09 <Tofe> so having CONFIG=debug wouldn't help much actually
Aug 30 11:31:46 <Tofe> and, thus, the flags I added in our bbappend
Aug 30 11:32:50 <Tofe> But I'll continue digging, this is probably a configuration that was never tested before by the Qt guys
Aug 30 11:35:24 <Herrie|Pre3> JaMa: OK
Aug 30 11:35:44 <Herrie|Pre3> My little experiment with adding /var/log/journal didn't work
Aug 30 11:36:53 <Tofe> Herrie|Pre3: ah, that's sad
Aug 30 11:37:08 <Herrie|Pre3> Let me find the log you guys might have a clue
Aug 30 11:37:18 <Herrie|Pre3> I.e. something with exists and volatile
Aug 30 11:38:02 <Tofe> Herrie|Pre3: in journald.conf you should have "auto" for the persistency
Aug 30 11:52:42 <Herrie|Work> Tofe: https://bpaste.net/show/e5a1fc3cf208
Aug 30 11:53:17 <Tofe> Herrie|Work: ah yes, you have to delete the /var/log symlink first
Aug 30 11:54:00 <Tofe> but I'm not sure where it is created in the first place
Aug 30 11:54:36 <Tofe> Herrie|Work: if we have an initrd for qemu, that may be a good place to do it, isn't it ?
Aug 30 12:16:53 <Herrie|Work> Tofe: It seems that the base-files are set by oe-core
Aug 30 12:16:58 <Herrie|Work> As volatile
Aug 30 12:21:42 <Herrie|Work> I guess I could do a .bb instead of .bbappend in our repo and then remove the volatile log from there just to get to the error
Aug 30 12:33:06 <Tofe> Herrie|Work: can't we just have an "install_append" that deletes it ?
Aug 30 12:34:24 <Herrie|Pre3> Tofe: Probably too
Aug 30 12:36:47 <Herrie|Work> Tofe: I couldn't get that to work. But for now I'll just try with a meta-webos-ports recipe where we just don't create it :P
Aug 30 12:53:36 <Tofe> ok :)
Aug 30 15:15:08 <Herrie|Pre3> Tofe: I'm toying a bit with the widevine & flash plugins a bit...
Aug 30 15:15:19 <Herrie|Pre3> Seems Flash doesn't work while it should :P
Aug 30 16:28:56 <Herrie|Pre3> Tofe: Can we pass Chromium parameters to luna-webappmgr for example?
Aug 30 16:29:08 <Herrie|Pre3> Or luna-qml-launcher?
Aug 30 16:29:48 <Tofe> which Chromium parameters ?
Aug 30 16:33:28 <Herrie|Pre3> Tofe: For widevine & flash we need to pass in some params it seems
Aug 30 16:33:47 <Tofe> build parameters ?
Aug 30 16:33:55 <Tofe> runtime parameters ?
Aug 30 16:34:03 <Herrie|Pre3> Like Flash version
Aug 30 16:34:09 <Herrie|Pre3> Runtime
Aug 30 16:36:26 <Tofe> you can do something like that: https://github.com/webOS-ports/luna-webappmanager/commit/2c496910c4dab8fbb9a6bdc68fa8121e6f823d49
Aug 30 16:36:57 <Herrie|Pre3> https://doc.qt.io/qt-5/qtwebengine-features.html#pepper-plugin-api
Aug 30 16:37:09 <Herrie|Pre3> Tofe: thnx
Aug 30 16:38:31 <Tofe> do we set "pluginsEnabled:true" in our qml code ?
Aug 30 16:38:54 <Tofe> because your link says that flash plugin should be active by default
Aug 30 16:40:10 <Tofe> is the libpepflashplayer.so present in our image ?
Aug 30 16:42:06 <Herrie|Pre3> Tofe: We do in the LunaWebEngineView.qml in LuneOS Components
Aug 30 16:42:23 <Herrie|Pre3> But there's also Settings in Browser app that stores these settings
Aug 30 16:42:28 <Herrie|Pre3> So not sure which one is used
Aug 30 16:42:56 <Herrie|Pre3> Doesn't seem there's debug output in WebEngineView to tell that plugins are enabled
Aug 30 16:43:22 <Herrie|Pre3> I took the .so from a recent ChromeOS ARM image and placed it on all locations mentioned in QT page :P
Aug 30 16:43:35 <Herrie|Pre3> So that should suffice
Aug 30 16:43:54 <Tofe> yes, so you can try with the command line, but I thnk the issue may be somewhere else
Aug 30 16:44:00 <Herrie|Pre3> However also read that Chromium 56 disables Flash by default. Could be QT somehow overlooked that
Aug 30 16:44:15 <Tofe> oh, ok, yes in that case...
Aug 30 16:44:36 <Herrie|Pre3> Since they introduced it with 5.7 which had Chromium 49 if I recall correctly where it was enabled by default still
Aug 30 16:44:53 <Herrie|Pre3> But flash should work when binary is there
Aug 30 16:45:13 <Tofe> I prefer to activate it explicitely anyway, this stuff just eats battery life
Aug 30 16:45:14 <Herrie|Pre3> I can confirm my configure of QtWebEngine shows it as enabled.
Aug 30 16:45:58 <Herrie|Pre3> I don't care too much about Flash as such. WideVine is bigger goals (i.e. Netflix, YouTube paid stuff etc ;)
Aug 30 16:46:14 <Herrie|Pre3> Which is really useful on tablet like TP :P
Aug 30 16:46:26 <Herrie|Pre3> But Flash should be easier to test
Aug 30 16:46:33 <Herrie|Pre3> But well I'll play around a bit
Aug 30 16:47:01 <Herrie|Pre3> Should have a new qemu image with /var/log/journal waiting at home to test as well
Aug 30 16:47:12 <Herrie|Pre3> Hopefully will be able to track qemu isssue
Aug 30 16:47:31 <Tofe> ok yes, widewine is a very nice to have
Aug 30 16:48:31 <Herrie|Pre3> Tofe: Not a fan of DRM but people consume media so we'd need to provide something for it ;)
Aug 30 16:48:57 <Herrie|Pre3> Legacy webOS never had proper support for it, so would be a nice feature :)
Aug 30 16:49:06 <Tofe> well, we can ship it in a separate .ipk later on, when we'll dominate the marker :)
Aug 30 16:49:09 <Tofe> market
Aug 30 16:49:23 <Herrie|Pre3> Tofe: Yup
Aug 30 16:50:02 <Herrie|Pre3> Seems WideVine might be compilable from source, not 100% sure. Flash is not but it's dying anyway
Aug 30 16:51:24 <Herrie|Pre3> They're available for ARMv7 anyway in some images :P
Aug 30 16:51:33 <Herrie|Pre3> So can always pull them from there and sideload
Aug 30 17:29:15 <Tofe> I think I found a convoluted way to fix the debug build... I couldn't identify the root cause though
Aug 30 17:29:43 <Tofe> (and the build error occurs at 96%, "of course!" )
Aug 30 17:30:29 <Herrie|2> Tofe S
Aug 30 17:30:31 <Herrie|2> :S
Aug 30 17:30:33 <Herrie|2> I hate that
Aug 30 17:32:40 <Tofe> https://paste2.org/2vNNKbZA
Aug 30 17:35:32 <Herrie|2> That works?
Aug 30 17:41:04 <Tofe> maybe :) my build is currently at 46%
Aug 30 17:48:57 <Herrie> Ah ok ;)
Aug 30 17:49:14 <Herrie> Trying to setup Qt here
Aug 30 17:49:19 <Herrie> First installing VS2016 Update 3
Aug 30 17:49:21 <Herrie> THen Qt
Aug 30 17:49:28 <Herrie> Hopefully Kits will automagically work
Aug 30 18:18:16 * Tofe is glad that the qtwebengine recipe build time is about the same as the diner cooking time
Aug 30 18:21:22 * Herrie is still waiting for pizza to be delivered, was suppossed to arrive 19:53 :P
Aug 30 19:21:32 <Tofe> my patch worked, it builds now. Now I have to check if the debug symbols are there...
Aug 30 19:22:35 <Tofe> yup, it improved the situation
Aug 30 19:30:13 <Herrie> :D
Aug 30 19:31:33 <Herrie> Seems TCL will revive the Palm brand finally in 2018
Aug 30 19:31:37 <Herrie> Curious what they'll bring
Aug 30 19:32:03 <Herrie> According to a NL Android site
Aug 30 19:32:33 <Herrie> https://www.androidplanet.nl/nieuws/palm-comeback-in-2018/
Aug 30 19:45:19 <Tofe> Herrie: interesting... but I fear it'll just be an Android device
Aug 30 19:52:18 <Herrie> Yup... But maybe hw kbd and unlocked bootloader
Aug 30 19:52:26 <Herrie> Unlike bb devices
Aug 30 20:03:13 <Tofe> In that case, we could do a port for that device eventually, as more as a symbol ;)
Aug 30 20:25:32 <Herrie> Let's see
**** ENDING LOGGING AT Thu Aug 31 03:00:01 2017