**** BEGIN LOGGING AT Fri Aug 18 03:00:02 2017
Aug 18 07:22:24 <Herrie|Pre3> Morning
Aug 18 08:09:26 <Herrie|Pre3> nizovn: ping
Aug 18 08:13:48 <nizovn> Herrie|Pre3: pong
Aug 18 08:14:54 <Herrie|Pre3> nizovn: I'm wondering how you build your webengine 5.9 for legacy. With meta-qt5 we cannot seem to get it to build :s
Aug 18 08:16:08 <nizovn> hm, i had no issues...
Aug 18 08:16:49 <Herrie|Pre3> You did clean build of just qt bits or you used yocto/oe?
Aug 18 08:17:35 <nizovn> yes, just qt from tarball sources
Aug 18 08:20:20 <Herrie|Pre3> nizovn: OK
Aug 18 08:23:49 <nizovn> seems you can build it manually as well - qmake is available under tmp-glibc/sysroots/x86_64-linux/usr/bin/qt5/qmake , but i'm not sure it's a good way
Aug 18 08:29:26 <Herrie|Pre3> nizovn: It's probably a silly typo somewhere in meta-qt5 just need to figure out wherre
Aug 18 08:29:43 <Herrie|Pre3> The GYP->GN probably has an issue somewhere
Aug 18 08:32:37 <nizovn> Herrie|Pre3: what error do you get?
Aug 18 08:35:18 <Herrie|Work> nizovn: https://bpaste.net/show/b1ea511255d2
Aug 18 08:52:52 <Herrie|Pre3> nizovn: Some digging suggests this might be caused by some cr/crlf or similar issue in a file :s
Aug 18 08:59:47 <nizovn> maybe it's python2 vs 3 problem
Aug 18 10:25:30 <Herrie|Pre3> I think I might have found a clue
Aug 18 10:26:19 <Herrie|Pre3> Found some "random" patch @ QT which supposedly fixes an issue with long paths in Yocto with Ninja
Aug 18 10:45:54 <Herrie|Pre3> Seems I get it to build now at least :P
Aug 18 10:46:08 <nizovn> :)
Aug 18 10:46:45 <Herrie|Pre3> It at least doesn't error out right away
Aug 18 10:47:13 <Herrie|Pre3> I would still need to add all our patches back but well :)
Aug 18 10:47:23 <Herrie|Pre3> Getting it to build is a good start already
Aug 18 12:10:38 <Herrie|Pre3> JaMa: Morning
Aug 18 12:10:47 <Herrie|Pre3> Seems that I can get 5.9 webengine to build now
Aug 18 12:10:56 <Herrie|Pre3> But lost VNC connection to home to verify
Aug 18 12:11:16 <Herrie|Pre3> I took latest recipe from qt's meta-qt5
Aug 18 12:11:24 <Herrie|Pre3> Where there are only 2 patches
Aug 18 12:13:18 <Herrie|Pre3> JaMa: Sent you the link by email.
Aug 18 12:15:50 <Herrie|Pre3> To fix toolchain and work around the path length
Aug 18 12:34:12 <JaMa> Herrie|Pre3: I've replied now, for what MACHINE are you building?
Aug 18 12:34:21 <JaMa> and do you have multilib toolchain on your host?
Aug 18 12:34:51 <JaMa> in my build with examples enabled I'm now stuck in qtdeclarative:
Aug 18 12:34:52 <JaMa> /OE/build/oe-core/tmp-glibc/work/i586-oe-linux/qtdeclarative/5.9.0+gitAUTOINC+96f6ba5f56-r0/git/tests/auto/qml/qqmlcomponent/tst_qqmlcomponent.cpp:586:10: fatal error: tst_qqmlcomponent.moc: No such file or directory
Aug 18 12:34:57 <JaMa>  #include "tst_qqmlcomponent.moc"
Aug 18 12:34:59 <JaMa>           ^~~~~~~~~~~~~~~~~~~~~~~
Aug 18 12:39:13 <Herrie|Pre3> JaMa: I'm trying Hammerhead
Aug 18 12:39:42 <Herrie|Pre3> I just have plain Ubuntu 16.04 and do our regular build as per instructions on wiki
Aug 18 13:41:40 <bshah> Tofe: you had EGL_BAD_DISPLAY error for while right? how you solved it?
Aug 18 16:21:01 <Tofe> Morning
Aug 18 16:21:50 <Tofe> bshah: that was, in my case, an issue with the andoid caf code
Aug 18 16:25:10 <Tofe> JaMa: moc should be generated from the .h parsing with "moc" tools, I guess. Maybe it's generated in an unexpected place, not reachable by the include statement ?
Aug 18 16:34:13 <JaMa> Tofe: it's some strange issue, after cleansstate it works, maybe related to issues Herrie  mentioned before
Aug 18 16:35:20 <Tofe> ok
Aug 18 16:35:20 <JaMa> remaining failures in my local build: meta-toolchain-qt5 nativesdk-packagegroup-qt5-toolchain-host nativesdk-qtwayland packagegroup-qt5-toolchain-target qt5-creator qtquick1 qttools qtwebengine qtwebview qwt-qt5
Aug 18 16:39:01 <Tofe> JaMa: not especially related to moc?
Aug 18 17:04:15 <JaMa> Tofe: it's related but more like some race-condition or improper cleanup of build directory when re-executing do_compile task
Aug 18 17:05:33 <Herrie> JaMa: I also have issues with QT/QML bits when I build for example Tenderloin after Hammerhead or vice versa
Aug 18 17:08:37 <JaMa> QT/QML bits include MOC?
Aug 18 17:11:27 <saidinesh5> JaMa: Qt bits can include moc. QML bits nope.
Aug 18 17:12:19 <saidinesh5> basically with moc you generate additional code for the classes in your .h files, so that they can work with Qt's signals and slots
Aug 18 17:19:06 <Tofe> saidinesh5, JaMa: yes, though it depends also what you are talking about with "QML"; a whole QML program/plugin can be a mix of QML and C++, and the latter will bring moc.
Aug 18 17:19:39 <saidinesh5> ditto
Aug 18 17:19:40 <saidinesh5> lol
Aug 18 17:19:54 <saidinesh5> okay time to start the halium download i think
Aug 18 17:20:01 <Tofe> :)
Aug 18 17:22:51 <saidinesh5> Tofe: out of curiosity, how much disk space is needed for halium?
Aug 18 17:23:21 <Tofe> we recommend at the very least 60GB
Aug 18 17:23:45 <Tofe> ah no wait sorry
Aug 18 17:23:54 <Tofe> halium, not luneos :p
Aug 18 17:24:28 <Tofe> I'd say 15GB ? I didn't really check, but halium is lighter than a whole ROM build
Aug 18 17:25:03 <Tofe> the tenderloin result I built was... 20MB :)
Aug 18 17:26:30 * saidinesh5 checks
Aug 18 17:26:35 <saidinesh5> well lune OS too..
Aug 18 17:30:29 <Tofe> let's give it a try with 80GB total, there shouldn't be any surprise. Less should be ok too, but I didn't check how much is really needed lately...
Aug 18 17:30:56 <Herrie> saidinesh: You should be OK with 60-80gb I'd say
Aug 18 17:31:01 <Herrie> Biggest eater is qtwebengine
Aug 18 17:32:08 <Tofe> yeah, its debug symbol alone is 1.5GB...
Aug 18 17:48:03 <saidinesh5> Ahh
Aug 18 17:48:14 <saidinesh5> you compile the whole thing from scratch for every device?
Aug 18 17:48:24 <saidinesh5> can''t i just reuse the rootfs done for say hammerhead?
Aug 18 17:48:32 <saidinesh5> (bad internet/electricity here today)
Aug 18 17:50:44 <Herrie> saidinesh5: In general yes, however with Yocto you can reuse certain bits
Aug 18 17:50:58 <Herrie> I.e. if both are armv7 etc you don't need to recompile everything ;)
Aug 18 17:51:00 * saidinesh5 wonders whether he should use his VPS for this then
Aug 18 17:51:08 <Herrie> So it reuses where possible
Aug 18 17:51:09 <saidinesh5> yeah both are msm8974s too actually
Aug 18 17:51:25 <Herrie> Especially useful when there's plenty of devices to build
Aug 18 17:52:13 <Herrie> saidinesh5: When you build for hammerhead you should be able to reuse quite some bits for your new target :D
Aug 18 17:52:22 <saidinesh5> Ahh cool
Aug 18 17:52:29 <Herrie> So you can start with building that
Aug 18 17:52:34 <saidinesh5> Aye
Aug 18 17:52:35 <Herrie> And then add your own target ;)
Aug 18 17:52:42 <Herrie> Yocto is pretty nice for this
Aug 18 17:52:44 <saidinesh5> so local build it is..
Aug 18 17:52:51 <Herrie> Not sure how OBS works for this
Aug 18 17:53:09 <Herrie> Yocto has a lot of backing from Intel for example
Aug 18 17:53:19 <Herrie> I like the approach in general
Aug 18 17:53:21 <saidinesh5> for sailfish it was pretty easy to reuse the  "rootfs" because all it was was  a list of rpms
Aug 18 17:53:38 <Herrie> Well with us also, however specific for each architecture
Aug 18 17:53:54 <Herrie> I.e. armv5, v6, v7, v8 would have different builds
Aug 18 17:53:57 <saidinesh5> yep
Aug 18 17:54:00 <Herrie> So would qemux86
Aug 18 17:55:14 <Herrie> Well in general with us it's a list of IPK's but the same principle
Aug 18 17:55:27 <Herrie> But we have qemux86 target as well for example
Aug 18 17:55:59 <saidinesh5> let me see if i can convince the guy to release a 3.10 kernel early for my device
Aug 18 17:56:04 <saidinesh5> that should save me a lot of pain..
Aug 18 18:01:22 <Herrie> Yeah :D
Aug 18 19:52:31 <Herrie> JaMa: I'm getting .moc failures at the examples now too :S
Aug 18 19:52:40 <Herrie> Any way to bypass building the examples?
Aug 18 19:52:44 <Herrie> Since we don't really need them
Aug 18 20:21:07 <JaMa> Herrie: for me it was resolved by cleansstate and rebuild
Aug 18 20:21:42 <JaMa> there is packageconfig in qtbase, but I think now they are enabled together with ptest
**** ENDING LOGGING AT Sat Aug 19 03:00:00 2017