**** BEGIN LOGGING AT Fri Aug 30 02:59:58 2013 Aug 30 08:00:07 JaMa: yeah saw that some days ago Aug 30 08:00:27 JaMa: maybe they are changing some parts of their configuration due to the last ddos attack Aug 30 08:27:11 panda-z_: ping Aug 30 08:27:36 morphis: pong Aug 30 08:27:49 panda-z_: you already had a chance to look at the device menu? Aug 30 08:29:13 morphis: yes, i've read the source code this days. should we implement statusbar first? Aug 30 08:29:36 panda-z_: I think both can happen independently Aug 30 08:30:05 I already have some small amount of work locally for a wifi status indicator Aug 30 08:31:14 morphis: ok, i have an idea. we can use Qt Quick Layouts from qtquickcontrols which is introduced from qt 5.1? Aug 30 08:31:32 yes Aug 30 08:31:48 we're going to add qtquickcontrols to meta-qt5 soon Aug 30 08:32:10 panda-z_: but afaik qt 5.1 is currently not supported in ubuntu Aug 30 08:32:31 http://blog.qt.digia.com/blog/2013/05/16/introducing-qt-quick-layouts/ Aug 30 08:32:44 seems like they are still at 5.0.2 Aug 30 08:34:00 panda-z_: what do you think about doing the device menu itself as standalone QML component? Aug 30 08:34:48 panda-z_: as we're currently doing a some rework in luna-next to extract the core UI part as shell Aug 30 08:35:07 so the user has the ability to use different shells which can implement different user interfaces Aug 30 08:35:14 morphis: you mean https://github.com/webOS-ports/luna-next-cardshell ? Aug 30 08:35:15 easy to implement but powerful feature Aug 30 08:35:17 yes Aug 30 08:35:43 the shells will base on a LunaNext QML plugin which is part of luna-next and gives them a solid base Aug 30 08:35:52 morphis: yeah, that sounds great Aug 30 08:35:54 so generic things should go into luna-next Aug 30 08:36:04 and I was thinking about putting the device menu there too Aug 30 08:36:37 so the core parts of the device menu are in luna-next and the visual parts of it are in luna-next-cardshell Aug 30 08:37:04 morphis: so any qml related work will be put into luna-next-cardshell, right? Aug 30 08:37:17 that depends Aug 30 08:37:50 as I said: we have a LunaNext plugin for QML which already contains some common functionalities Aug 30 08:38:10 and will also contain the wifi indicator for example as thats nothing every shell needs to reimplement Aug 30 08:39:21 and some parts of the device menu could also be part of the LunaNext plugin Aug 30 08:39:39 panda-z_: it's up to you how it suites best but thats at least the idea I have Aug 30 08:40:45 morphis: so how can i test my changes to luna-next-cardshell ？ in qt creator ? Aug 30 08:40:54 yes Aug 30 08:41:05 there is a qml project and a stub for the LunaNext plugin Aug 30 08:41:27 morphis: ok, that's great Aug 30 08:42:10 panda-z_: just start with the cardshell as base and we can look later what we can move out of it Aug 30 08:42:33 I think I can also play it in virtualbox, but currently the box image seems broken Aug 30 08:42:43 yes Aug 30 08:42:56 JaMa removed some changes from master on last rebase Aug 30 08:43:09 panda-z_: but they are back soon Aug 30 08:43:33 panda-z_: and also it will be easier to test the shell implementation as we're not packing it into the binary with the use of a resource file anymore Aug 30 08:44:02 you can just pack all the qml files in /usr/palm/luna-next/shells/card and restart luna-next Aug 30 08:45:20 morphis: ha, it's so convenient Aug 30 08:45:28 :) Aug 30 08:46:47 morphis: I'm also build my own image with bitbake and met a problem with opkg-cl, do you have some ideas? logs will be paste lately Aug 30 08:47:03 panda-z_: can be some dependency problem Aug 30 08:47:32 | + opkg-cl -f /home/panda/webos-ports/webos-ports-env/webos-ports/tmp-eglibc/work/qemux86-webos-linux/webos-ports-dev-image/1.0-r0/opkg.conf -o /home/panda/webos-ports/webos-ports-env/webos-ports/tmp-eglibc/work/qemux86-webos-linux/webos-ports-dev-image/1.0-r0/rootfs --force_postinstall --prefer-arch-to-version update Aug 30 08:47:32 | Aug 30 08:47:32 | /home/panda/webos-ports/webos-ports-env/webos-ports/tmp-eglibc/work/qemux86-webos-linux/webos-ports-dev-image/1.0-r0/temp/run.do_rootfs.16077: line 262: 16302 Segmentation fault (core dumped) opkg-cl -f $INSTALL_CONF_IPK -o $INSTALL_ROOTFS_IPK --force_postinstall --prefer-arch-to-version update Aug 30 08:47:59 where is the error? :) Aug 30 08:48:06 ah segmentation fault Aug 30 08:48:23 thats quite interesting and hard to trace Aug 30 08:48:34 panda-z_: you retried to build the image? Aug 30 08:49:24 morphis: yes, should i do a clean build? Aug 30 08:49:48 no, just retry to build the image Aug 30 08:50:41 morphis: have retried several times, same error Aug 30 08:51:19 JaMa: any idea? Aug 30 08:51:44 ERROR: Task 7 (/home/panda/webos-ports/webos-ports-env/webos-ports/meta-webos-ports/recipes-core/images/webos-ports-dev-image.bb, do_rootfs) failed with exit code '1' Aug 30 08:52:21 panda-z_: how buch memory your build machine has? Aug 30 08:53:41 morphis: 8 GB Aug 30 08:54:02 should be enough :) Aug 30 08:55:01 yes, segfault error is difficult to trace, though Aug 30 08:56:21 yeah Aug 30 08:56:28 maybe JaMa saw something similar already Aug 30 08:56:39 he's more the guy for such build things Aug 30 08:58:04 here's the entire log: http://pastebin.com/pTyGWNnf Aug 30 09:00:02 panda-z_: with the device menu work: the currently top most priority is to get something visible and looking complete the integration with the several system services comes later Aug 30 09:00:27 we need to show luna next soon to tell the community what we have and that the vision is valid and a way to succeed Aug 30 09:01:45 morphis: ok Aug 30 09:03:53 panda-z_: I'm not sure why only some build trigger this segfault, but I think I know what it is, mmt Aug 30 09:04:21 morphis: panda-z_: http://git.openembedded.org/openembedded-core/commit/?id=e4190b601c0212f0009ddb2dfa00d2880bddbc29 Aug 30 09:04:47 exclude lists were added to oe-core lately and we're missing this fix in our webos-ports/master I guess, checking now Aug 30 09:05:01 ack Aug 30 09:05:06 JaMa: thanks Aug 30 09:05:28 I'll cherry-pick it to oe-core instead of merging new stuff Aug 30 09:05:33 ok Aug 30 09:06:04 JaMa: thanks, will cherry-pick it and have a try Aug 30 09:06:10 panda-z_: upgrade to latest webos-ports/master (instead of locked revision from layers.txt) and confirm it works Aug 30 09:06:20 or cherry-pick it yourself, right :) Aug 30 09:07:39 morphis: last rebuid requires nyx-modules and serviceinstaller refresh (from new versions from upstream, can you look at that)? Aug 30 09:07:42 /OE/webos/owpb/webos-ports/meta-webos-ports/recipes-webos/serviceinstaller/serviceinstaller.bb, do_patch Aug 30 09:07:45 /OE/webos/owpb/webos-ports/meta-webos-ports/recipes-webos/nyx-modules/nyx-modules.bb, do_configure Aug 30 09:08:23 nyx-modules rebase should be enough | Component version from build system (5.2.0) != configured version (5.1.0) Aug 30 09:09:05 JaMa: ok Aug 30 09:09:16 JaMa: the patch for serviceinstaller we have doesn't apply anymore? Aug 30 09:09:44 yes, I can update this one Aug 30 09:10:31 JaMa: can you try nyx-modules with 5bdd7da91a178954e500d6872045b2933d369f29? Aug 30 09:10:43 pushed it as webOS-ports/master-next rebased on latest upstream Aug 30 09:10:55 trying Aug 30 09:10:57 if it works I will push it as webOS-ports/master Aug 30 09:11:52 will report result when build is finished, have to leave now, bye Aug 30 09:13:00 morphis: EricBlade made the same patch in serviceinstaller, so I'll drop ours Aug 30 09:13:09 JaMa: great Aug 30 09:16:35 build running, if it works I'll push both Aug 30 09:16:55 morphis: any idea about that cmake error? Aug 30 09:17:11 JaMa: you mean the cmake one? Aug 30 09:17:40 s/cmake/qmake/ Aug 30 09:18:23 broken cmake build with cmake changes from qtbase Aug 30 09:20:23 yes, not really Aug 30 09:20:47 either we need to export QT_.._HOST_BINS or we replace the variable with the explicit path Aug 30 09:21:18 but that still doesn't solves the problem that it seems like we're having a direct dependency on qtbase here Aug 30 09:21:42 as why is cmake complaining about it at all when not building against qt? Aug 30 09:22:48 I was thinking about some weak assignment in cmake file (like bitbake does), but that doesn't work well because it would hardcode the path to sysroot Aug 30 09:24:05 nyx-modules need a bit more work: Aug 30 09:24:06 | /OE/webos/owpb/webos-ports/tmp-eglibc/work/qemux86_64-webos-linux/nyx-modules/5.2.0-90+wop+gitrAUTOINC+5bdd7da91a-r0/git/src/tests/test-touchpanel.c:58:70: error: 'this' undeclared (first use in this function) Aug 30 09:24:19 JaMa: ok, will rework it later Aug 30 09:25:21 JaMa: hm Aug 30 09:32:55 JaMa: but I am still asking me why is cmake evaluating the path Aug 30 09:33:17 as building cmake itself should be infected by any file in this way when installed in sysroot Aug 30 09:33:32 I'll try to reproduce it on my machine to see which part of configure is calling that Aug 30 09:33:53 s/should/shouldn't/g ? Aug 30 09:35:09 ok Aug 30 09:51:32 I'll temporary disable test-touchpanel in my local build to test other changes from rebase Aug 30 13:40:34 morphis: do you also see issues booting qemux86* image on first boot? 2nd seems to be working better Aug 30 13:52:39 JaMa: only that luna-next doesn't come up because of the missing mesa changs Aug 30 14:01:54 ok, in my build it was login.service Aug 30 14:02:15 now I'm debugging mesa too, because it was built with your 2 patches but still missing llvmpipe Aug 30 14:02:40 bitbake -e looks correct so it could be another dependency missed when added through PACKAGECONFIG Aug 30 14:19:15 hm, I had to get the packages back from the dark grounds of my git repository as they were gone form morphis/work and master so I may have missed something Aug 30 14:19:23 as I had llvmpipe compiled in Aug 30 14:21:04 and it was working in virutalbox with a qemux86 image Aug 30 14:34:41 morphis: I've grepped whole reflog and found them only in latest cherry-pick from morphis/work, so for some reason I never had them in my master branch :/ Aug 30 14:35:00 but it looks that after enabling gallium-gbm I have llvm installed again Aug 30 14:36:14 I promise to tag rebased masters (with date for now) before pushing them Aug 30 15:24:51 hi Aug 30 15:51:02 morning Garfonso Aug 30 15:52:52 how are things moving along? Aug 30 15:53:02 I lost track a bit, on the broader scale.. Aug 30 20:51:33 morning ! Aug 30 20:52:38 JaMa: your recent talk doesn't encourage me to update my master branch ;) Aug 30 20:54:10 Tofe: it's not about current master but master-next Aug 30 20:54:29 but more change will come soon and invalidate many checksums again Aug 30 20:55:36 :)great Aug 30 21:45:59 Roy__away: How's that blog update coming along ;) Been 6 months now ;) Aug 31 01:21:23 Herrie_zZzZzz: I know, I know! **** ENDING LOGGING AT Sat Aug 31 02:59:58 2013