Re: Review Request 128002: address alignment issue with the OS X native "macintosh" style
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/128002/#review95755 --- Ship it! Also the navigator got fixed by you after all. Thanks, René! - Marko Käning On May 24, 2016, 2:50 p.m., René J.V. Bertin wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/128002/ > --- > > (Updated May 24, 2016, 2:50 p.m.) > > > Review request for KDE Software on Mac OS X and KDE Frameworks. > > > Repository: kio > > > Description > --- > > This patch addresses a misalignment issue in KUrlNavigator when using the > native "macintosh" widget style on OS X. > > Cf. https://bugs.kde.org/show_bug.cgi?id=296845 > > > Diffs > - > > src/filewidgets/kurlnavigatorbuttonbase.cpp 8b06bfa > > Diff: https://git.reviewboard.kde.org/r/128002/diff/ > > > Testing > --- > > On OS X 10.9 with frameworks 5.20.0 (sic), Qt 5.6.0 using my MacPorts build. > > The additional attribute doesn't have any effects for me with other styles, > but this still has to be verified on other platforms. > > > File Attachments > > > The widget concerned in Kate's interface, showing the result with and without > the patch. It is clear which is which. > > https://git.reviewboard.kde.org/media/uploaded/files/2016/05/24/ca3c7872-68ee-4ed6-b1f4-4d0b05111be1__Screen_Shot_2016-05-24_at_14.37.43.png > other widgets remain to be done > > https://git.reviewboard.kde.org/media/uploaded/files/2016/05/24/71b20243-8dda-4f22-89e1-21559b767fd5__Screen_Shot_2016-05-24_at_14.38.30.png > > > Thanks, > > René J.V. Bertin > > ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 128005: fix alignment issue with the OS X native "macintosh" style
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/128005/#review95754 --- Ship it! Hi René, nice to see that you finally found a solution for this old problem! Congrats, Marko - Marko Käning On May 24, 2016, 6:39 p.m., René J.V. Bertin wrote: > > --- > This is an automatically generated e-mail. To reply, visit: > https://git.reviewboard.kde.org/r/128005/ > --- > > (Updated May 24, 2016, 6:39 p.m.) > > > Review request for KDE Software on Mac OS X, KDE Frameworks and Christoph > Feck. > > > Repository: kwidgetsaddons > > > Description > --- > > This patch fixes the size & alignment issue discussed in > https://bugs.kde.org/show_bug.cgi?id=296810 > > > Diffs > - > > src/kmultitabbar.cpp f7739fa > > Diff: https://git.reviewboard.kde.org/r/128005/diff/ > > > Testing > --- > > On OS X 10.9.5 with KF5 5.20.0, Qt 5.6.0 and Kate 16.4.01 . No adverse > effects when used with other themes on this platform. > > > File Attachments > > > Kate5 running with the native "macintosh" qpa & theme, showing correct > alignment. > > https://git.reviewboard.kde.org/media/uploaded/files/2016/05/24/fe90cba8-f6de-4e5c-8732-4cbebf4abe8b__Screen_Shot_2016-05-24_at_18.38.16.png > > > Thanks, > > René J.V. Bertin > > ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: [KDE/Mac] Question about goal of Windows/Mac frameworks
Hi René, On 22 Oct 2015, at 19:24 , René J.V. Bertinwrote: > Not exactly, no. The patch was rejected in the presented form, but we were > invited to file a bug report (I think it's > https://bugreports.qt.io/browse/QTBUG-44473). Here we (me, IIRC) managed to > make the case for special needs by MacPorts and family. > My understanding is that the door is still open for a QSP patch that does not > alter the default QSP behaviour, which is exactly what my patch does (i.e. it > requires activation). ok, you're right. >> Could it be, that we have to introduce a QSP patch in MacPorts’ qt5-mac to >> revert back to the Linuxy way of XDG paths? > > Have you forgotten I'm working on that, and that I actually created a qt5-kde > port so that (y)our KF5 ports could depend on a port with the required > patch(es) and install layout? How could I forget that? ;-) I was merely trying to trigger the discussion! Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: [KDE/Mac] Question about goal of Windows/Mac frameworks
Thanks René and Jeremy, On 22 Oct 2015, at 22:43 , Jeremy Whitingwrote: > ...It sounds like a good solution for embedding a copy of Qt next > to each application for windows use (and maybe for osx use too if > resources don't make it completely unneccessary), but not for the > macports shared Qt case. for clarifying this. Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: [KDE/Mac] Question about goal of Windows/Mac frameworks
Hi Ralf, On 22 Oct 2015, at 08:35 , Ralf Habackerwrote: > umbrello for example depends on about 50 other libraries and packages > https://build.opensuse.org/project/show/home:rhabacker:branches:windows:mingw:win32:KF511. > Not patching Qt requires to repack every single package :-(, by either > hacking the cmake build system to use different install locations or to > reorder the installed files after cmake installing. > > Having Qt support for "standard linux path layout" for example by > extending qt.conf to support QStandardPath (qt.conf is already required > for KDE on Windows) shortcuts this repackaging need completely. have you followed the discussion with Qt's developers regarding the QSP patch [1]? If not, I advise you to do a little reading there! Qt won’t ever support such an approach, i.e. one would have to patch it, if KDE itself doesn’t come with its own provisions for this... If every single KDE application wants to be self-contained - to be more easily distributable - then that’s fine, especially for bigger apps like KMail, DigiKam, Marble, KDEnlive… This would perhaps even make a distribution via the App Store possible. ;-) However, if one wants to avoid all the duplication of libs to be shipped, then one better uses a package management system like MacPorts, Homebrew, Fink on OSX and who knows what on Windows. This however will require extra efforts for those systems, if KDE doesn’t somehow envisions such distribution mechanisms for non-Linux distros. Wondering where things are heading to now... I think it’s great, that this thread(s) started a discussion about this pressing topic. :-) Could it be, that we have to introduce a QSP patch in MacPorts’ qt5-mac to revert back to the Linuxy way of XDG paths? Greets, Marko [1] https://codereview.qt-project.org/#/c/103277/ ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Version numbering at download location
Hi David, On 10 Oct 2015, at 15:37 , David Faurewrote: > OK, so we're not talking about the organization of the FTP directories, but > about HTML pages. > That's quite different (a single HTML page could point to multiple FTP > directories, for instance). yep. > Does https://www.kde.org/info/kde-frameworks-5.15.0.php work for this? > Well, the above has all you need for z==0. Looks good for z=0… > For z > 0, I previously sometimes made a separate page, but that's not > automated and I'm lazy > so I actually forgot for kservice 5.14.z. > Might be simpler for me to just add entries into that table above, and then > it would work better > for you, right? I think it’s indeed simpler to just append a section to above page which lists all incoming patch releases for the respective x.y version. This way MacPorts’ livecheck could easily find out whether there is a new patch-version z present. The creation of the patch-release section could surely be automated by a script which would have to be executed on every patch release. Thanks, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Version numbering at download location
On 05 Oct 2015, at 10:08 , David Faure <fa...@kde.org> wrote: > On Monday 05 October 2015 09:34:04 Marko Käning wrote: >> The issue is that the version regexp match should preferably >> happen on a single web page. With this setup the full version numbers are >> spread over two dir >> levels… :-( > > I'm not sure I understand which web page this is about. > I can do things differently for hotfix releases on www.kde.org, but maybe > this isn't what you meant? In MacPorts the version check is usually done by downloading the web-page where the corresponding source code archives can be grabbed from. This is usually a simple web-page created by source-forge, github, bitbucket or the like… Such a HTML page would contain all valid version numbers (together with other infos) of a given code-archive. The regexp to be defined now grabs the version numbers in the desired format, i.e. x.y.z and thus allows to cross-check all numbers found against the one currently in use be the installed version of the corresponding software. Yet, KF5’s approach means that you have at first a page where you get a choice between x.y’s and then at the following page - a lever lower - you only start to see x.y.z refinements... Does KDE perhaps have another overview web-page where for every code-archive distributed all currently available versions are listed?! That would come very handy for me now! ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Building KF in version 5.14 on OSX fails for framework kdesignerplugin
Hi David, now that I learnt, that we can have bug-fix releases in already released KF5 versions, I’d suggest a 5.14.1 for the kdesignerplugin, which contains Kurt’s fix: On 18 Sep 2015, at 20:42 , Marko Käning <mk-li...@mailbox.org> wrote: > Dear Kurt, > > On 16 Sep 2015, at 15:20 , Kurt Hindenburg <kurt.hindenb...@gmail.com> wrote: >> Git commit 6cbf534c3cd2c5bdda819e092c9778c27e814161 by Kurt Hindenburg. > > thanks for your commit! > > Looking forward to release 5.15 then, David. > :) That would finally allow me to test the 5.14 on a MacPorts-based OSX VM. What do you think about that? Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Building KF in version 5.14 on OSX fails for framework kdesignerplugin
Well, I know, I could simply make my Portfile use a patch file… On 05 Oct 2015, at 21:18 , Marko Käning <mk-li...@mailbox.org> wrote: > I’d suggest a 5.14.1 for the kdesignerplugin, which contains Kurt’s fix: … or simply wait a a tiny bit longer… Perhaps that’s less hassle then you trying to do create a another tiny release step for this framework given that you’ll come up within the next few days with new version 5.15.0 anyways! :) ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Version numbering at download location
Hi David, On 05 Oct 2015, at 09:27 , David Faurewrote: > Yes, this fix allows to have hotfix releases of single frameworks. Look at > KService 5.14.1, 5.14.2 and 5.14.3. > They are in the 5.14 subdir, where they belong. ok, I wasn’t aware of that. > x.y.z where z is > 0 is *always* about single framework hotfixes, never about > all frameworks, > hence the new subdir numbering scheme. Understood. > I'm sure you can remove "everything after the last dot" in your regexp, > otherwise I can > help with that :-) The regexp isn’t the point! ;-) The issue is that the version regexp match should preferably happen on a single web page. With this setup the full version numbers are spread over two dir levels… :-( Need to check whether this can be achieved anyway within MacPorts Portfile syntax by some Tcl magic. Thanks for the pointer! Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Version numbering at download location
Hi David, I was surprised to find that up to version 5.3.0 you were still using the trailing zero for the folder name below http://download.kde.org/stable/frameworks/ but from 5.4 on you started to omit it. The archives in those folders but still have that trailing zero... Does this mean that there will never be a x.y.z folder ever again on that page, or is that only temporally so? I am thinking about formulating the right regex for the version livecheck for MacPorts’ kf5-* ports, which is why I am asking. To be honest, I’d prefer, if you could revert back to x.y.z, as this simplifies the setup of the version livecheck tremendously! Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Building KF in version 5.14 on OSX fails for framework kdesignerplugin
Dear Kurt, On 16 Sep 2015, at 15:20 , Kurt Hindenburgwrote: > Git commit 6cbf534c3cd2c5bdda819e092c9778c27e814161 by Kurt Hindenburg. thanks for your commit! Looking forward to release 5.15 then, David. :) Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Building KF in version 5.14 on OSX fails for framework kdesignerplugin
Hi kdesignerplugin devs, On 14 Sep 2015, at 15:00 , Kurt Hindenburgwrote: > > It appears the QT_VERSION check in d53ec9b97d7b353ea4cce54d5ecae3c93b933ddd > is not enough when using Qt 5.4.x anyone out there who can fix kdesignerplugin regarding this version check so that the next KF5 release is build-able again on Qt 5.4.x? Greets, Marko P.S.: Thanks, Kurt, for investigating this! :) ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Building KF in version 5.14 on OSX fails for framework kdesignerplugin
Hi devs, the other day I had built version 5.13 successfully on a OSX/MacPorts-based VM. And so I wanted to try the same with the newest 5.14 just now, but failed! :( Any idea why those three interfaces are undefined? Greets, Marko --- :info:build [ 85%] Generating qrc_testplugin.cpp :info:build cd /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests && /opt/local/libexec/qt5-mac/bin/rcc -name testplugin -o /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests/qrc_testplugin.cpp /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/kdesignerplugin-5.14.0/autotests/testplugin.qrc :info:build /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/src/kdewebkitwidgets.cpp:22: Error: Undefined interface :info:build [ 85%] Generating testpluginwidgets.cpp :info:build cd /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests && ../src/kgendesignerplugin -o /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests/testpluginwidgets.cpp /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/kdesignerplugin-5.14.0/autotests/testplugin.widgets :info:build make[2]: *** [src/kdewebkitwidgets.moc] Error 1 :info:build make[2]: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build make[1]: *** [src/CMakeFiles/kdewebkitwidgets.dir/all] Error 2 :info:build make[1]: *** Waiting for unfinished jobs :info:build /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests/minimalwidgets.cpp:23: Error: Undefined interface :info:build make[2]: *** [autotests/minimalwidgets.moc] Error 1 :info:build make[2]: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build make[1]: *** [autotests/CMakeFiles/minimalplugin.dir/all] Error 2 :info:build [ 86%] Generating testpluginwidgets.moc :info:build cd /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests && /opt/local/libexec/qt5-mac/bin/moc @/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests/testpluginwidgets.moc_parameters_debug :info:build /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/autotests/testpluginwidgets.cpp:25: Error: Undefined interface :info:build make[2]: *** [autotests/testpluginwidgets.moc] Error 1 :info:build make[2]: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build make[1]: *** [autotests/CMakeFiles/testplugin.dir/all] Error 2 :info:build /opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build/src/kdewidgets.cpp:69: Error: Undefined interface :info:build make[2]: *** [src/kdewidgets.moc] Error 1 :info:build make[2]: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build make[1]: *** [src/CMakeFiles/kf5widgets.dir/all] Error 2 :info:build make[1]: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build make: *** [all] Error 2 :info:build make: Leaving directory `/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build' :info:build Command failed: cd "/opt/local/var/macports/build/_Users_marko_WC_GIT_macports-kde_dports_kde_kf5-kdesignerplugin/kf5-kdesignerplugin/work/build" && /usr/bin/make -j5 -w all VERBOSE=ON :info:build Exit code: 2 :error:build org.macports.build for port kf5-kdesignerplugin returned: command execution failed :debug:build Error code: CHILDSTATUS 4898 2 ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 123075: do not require X11 on Mac OS X
On March 20, 2015, 8:07 a.m., Martin Gräßlin wrote: as in other similar requests: -2 from my side Martin Gräßlin wrote: To extend: I think the way is wrong. If it now builds on MacOS the required is wrong. It should be an optional find_package properly ifdefed. Christoph Cullmann wrote: Actually, you don't want that it is optional as you really don't want that it ever is found on MacOS. If you install an XQuartz for legacy apps, it will be found, and you will have a completly mess as result ;=) Martin Gräßlin wrote: Christoph, that argument is wrong. The same would be the case with Windows and any other platform which optionally offers X11 (this includes Linux!). CMake can handle the situation quite well to disable an unwanted build dependency. If a user installs XLib headers (which is not the same as just installing XQuartz) we should expect the user to disable the cmake build option. Christoph Cullmann wrote: That means that nobody will get a senseful build on Mac, if he doesn't disable that optional dependency. Thats the opposite of a normal optional dependency, that leads to missing features if not found but not complete mess if found. I tried to compile frameworks stuff on Mac, and IMHO, really, that makes it close to unusable, that you need to tweak each cmake call just to have something usable, if you have too much stuff installed. I never had to do that on Linux/Other Unices. Martin Gräßlin wrote: That means that nobody will get a senseful build on Mac What since when is XLib as a build dependency available by default on OS X? Christoph Cullmann wrote: Actually, if you work with Frameworks there, you will install something over MacPorts or Homebrew, and yes, you will have XQuartz installed, to run some legacy apps and there is really no user understandable way to install XQuartz without its devel headers, the .dmg will just install everything, I was suprised myself that I have X devel headers just by installing an X-Server. My first workaround was just to deinstall XQuartz, later I started to tweak CMake options or do exactly the same fixes as above. And yes, I think, per default, without any magic options, frameworks should just build to a usable state. And I see 0 reason to ever have even an optional with x11 build of an frameworks application. But I might be wrong. Therefore I don't vote here or say ship it, just wanted to state my concerns ;=) Martin Gräßlin wrote: my concern is that we make our CMakeLists.txt way more complex (it's not the only framework which recently saw such a proposed change) and work around broken systems in our CMakeLists. That's something I do not want to see - if the downstream packaging sucks, it needs to be fixed there. We would tell the same to every Linux distribution. I do not want to see such OS specific changes go in as X11 is not OS specific - all operating systems we support in KDE can provide X11 and on all OS there are alternative windowing systems. What I don't want to see is something like: if (NOT APPLE AND NOT WINDOWS AND NOT LINUX_ANDROID AND NOT LINUX_UBUNTU_PHONE AND NOT LINUX_SAILFISH AND NOT LINUX_WEBOS AND NOT ... if we start with one platform, where do we end? Is adding the check for Apple OK and Windows not? Christoph Cullmann wrote: We could wrap the X11 search in a own module that will not do anything on Apple, to avoid the if stuff. On the other side, even with the if, it still avoids that we have some combinations feasible in the frameworks, that we need will not test anyway, e.g. apple + X11. That avoids complexity, too. Actually I would be OK do have the same for WIN, too, yeah, still better than a code path that you can configure in, that actually will not work. Martin Gräßlin wrote: I'm not worried about the code paths - everything X11 specific should be runtime wrapped anyway in a if (QX11Info::isPlatformX11()). If not: that needs fixing. What's the status of this one? - Marko --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123075/#review9 --- On March 19, 2015, 11:59 p.m., Harald Fernengel wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123075/ --- (Updated March 19, 2015, 11:59 p.m.) Review request for KDE Frameworks and Michael Palimaka. Repository: kdesu Description --- do not require X11 on Mac OS X Diffs - CMakeLists.txt 9623483d6f11f9cdb7d7dc19decfd7cf5e86d079 Diff: https://git.reviewboard.kde.org/r/123075/diff/ Testing ---
Re: Review Request 123692: autotests: Use QTEST_GUILESS_MAIN
On June 13, 2015, 10:26 a.m., Heiko Becker wrote: Ping? Hi Heiko, I've added this to our KDE's Phabricator task [T398](https://phabricator.kde.org/T398). ATM I can't test on my own OSX/CI system, as that is in limbo since we changed to Scarlett's new KDE/CI system. But I'll come back to this RR, once I had a chance to check it independently on OSX. - Marko --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123692/#review81439 --- On May 8, 2015, 9:15 p.m., Heiko Becker wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123692/ --- (Updated May 8, 2015, 9:15 p.m.) Review request for KDE Frameworks and Jeremy Whiting. Repository: knewstuff Description --- The tests still work fine with it, allowing them to run without a display server. Diffs - autotests/knewstuffauthortest.cpp 5ca7c4903e90a5b8bf377c3c12d5bbe7c0623002 autotests/knewstuffentrytest.cpp 7862ebbbf7c6a913cb4d6d0b70b2edf151ec6327 Diff: https://git.reviewboard.kde.org/r/123692/diff/ Testing --- Built and ran the tests successfully. Thanks, Heiko Becker ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: A CI dashboard with multiple versions of Qt5 on Linux
Hi Jan, On 12 Jun 2015, at 01:17 , Jan Kundrát j...@kde.org wrote: http://ci-logs.kde.flaska.net/matrix.html that looks really neat, indeed! :-D Congratulations, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: missing dep on OSX CI?
Hi David, On 06 Jun 2015, at 10:37 , David Faure fa...@kde.org wrote: KActivities fails with the error below. Is boost::optional installed? 04:17:31 /Users/jenkins/builds/kactivities/kf5-qt5/src/utils/continue_with.h:14:10: fatal error: 'boost/optional.hpp' file not found 04:17:31 #include boost/optional.hpp yes, it is installed, but below /opt/local! I’ve added you to https://phabricator.kde.org/T353 Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: CI for kf5-qt5 is not green anymore
Hi Scarlett, I’ve logged onto my VM as user jenkins, i.e. a GUI session is now available. So you could run a test-build for a project which failed before on you. Let me know how it goes. Greets, Marko signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 123595: Fix KUser test for Mac.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123595/#review79919 --- Looks like the test is still failing: https://build.kde.org/job/kcoreaddons%20master%20kf5-qt5/24/PLATFORM=OSX,compiler=clang/testReport/(root)/TestSuite/kusertest/ Anything I could do? - Marko Käning On May 2, 2015, 8:45 p.m., David Faure wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/123595/ --- (Updated May 2, 2015, 8:45 p.m.) Review request for KDE Frameworks and Marko Käning. Repository: kcoreaddons Description --- According to CI [1], an invalid user belongs to nogroup on Mac. Not sure if this is true on all OSX installations though? https://build.kde.org/view/Frameworks%20kf5-qt5/job/kcoreaddons%20master%20kf5-qt5/PLATFORM=OSX,compiler=clang/19/testReport/%28root%29/TestSuite/kusertest/ Diffs - autotests/kusertest.cpp d17a2d3e97d5056524281eb18766377e48a0da35 Diff: https://git.reviewboard.kde.org/r/123595/diff/ Testing --- Still passes on Linux; should fix the CI for mac, AFAICS. Thanks, David Faure ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: stable-KF5-QT5 QT version...
Hi Scarlett, On 18 Apr 2015, at 17:17 , Scarlett Clark sgcl...@kubuntu.org wrote: 15:09:51 with requested version 5.4.0”. this is not a matter of whether apps use stable-kf5-qt5 or kf5-qt5 branch groups, it is just about the notion devs have about requirements to be expected for stable-kf5-qt5. And this is clearly not the CI team’s job to get this fixed. ;) Many believe that 5.3.2 isn’t being developed for anymore and thus simply assume 5.4.0 to be the minimum requirement, also on stable-kf5-qt5! This is - of course - not in line with out CI system(s) and has popped up a couple of times on K-F-D in the last half year... So, I guess, we all need to come to a decision about what stable-kf5-qt5 actually means for all of KDE! Greets, Marko signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kdepimlibs fails to build on branch master
Hi Daniel, On 12 Apr 2015, at 09:50 , Daniel Vrátil dvra...@kde.org wrote: it's referring to a file in kdepim-runtime.git/resources (where Knut originally was before it was moved to kdepimlibs auto tests). I see. I don't understand Mac much, but is the file actually needed for the Knut resource? Well, the build fails because of it, thus it’s needed for that at least. :) Whether it is still really required I have no clue. It's installed, but it's only used for unit tests. Yep, and those shall be run. If it is still needed, we can just copy the file from kdepim-runtime.git to the knut folder in kdepimlibs and adapt the CMakeLists.txt. Go for it, please! :-) Thanks, Marko signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
New KDE/CI: qca fails to build on branch qt5 for branch group kf5-qt5 on platform OSX
https://build-sandbox.kde.org/job/qca%20qt5%20kf5-qt5/PLATFORM=OSX,compiler=clang/7/console ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kdepimlibs fails to build on branch master
Here is now the detailed information available from the new CI system: https://build-sandbox.kde.org/job/kdepimlibs%20master%20kf5-qt5/9/PLATFORM=OSX,compiler=clang/console Smells like a real Qt bug, indeed. signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120649: Encode the URIs which end up in DTD files
On Oct. 24, 2014, 12:05 a.m., Marko Käning wrote: I will have to test this in a specially configured OSX/CI VM. Will report back on it. Herewith I can report, that on Scarlett's new CI system we were able to verify that the file ```/opt/kde/install/darwin/mavericks/clang/kf5-qt5/frameworks/kdelibs4support/inst/Library/Application\ Support/kf5/kdoctools/customization/dtd/kdex.dtd``` correctly contained the escaped space in ```Application Support```! - Marko --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120649/#review69058 --- On Feb. 28, 2015, 11:05 p.m., Luigi Toscano wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120649/ --- (Updated Feb. 28, 2015, 11:05 p.m.) Review request for Build System, KDE Software on Mac OS X, KDE Frameworks, and kdewin. Repository: kdelibs4support Description --- The URI need to be encoded, because some valid characters for filenames are not valid according RFC 2396. Easy way to trigger the issue: when the path contains spaces, as it happens on MacOSX builds. The function is duplicated from the matching functions in kdoctools because: - this module will go away in the not far future; - the function is not generic enough to be used outside (yet) See also https://git.reviewboard.kde.org/r/120648/ for the twin review on kdoctools. Diffs - cmake/uriencode.cmake PRE-CREATION src/CMakeLists.txt 4a1d80d Diff: https://git.reviewboard.kde.org/r/120649/diff/ Testing --- It compiles, but I can't properly test Mac and Windows scenarios Thanks, Luigi Toscano ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kdepimlibs fails to build on branch master
Hi folks, after a long time I am finally able to build kmime [1] and thus also kimap and kmbox needed for kdepimlibs! :-D Now I can tackle kdepimlibs again, but the build fails because of a missing template file: --- -- Configuring done CMake Error: Target akonadi_knut_resource Info.plist template /Users/marko/WC/KDECI-builds/kf5-qt5/kdepimlibs/akonadi/autotests/testresource/../Info.plist.template could not be found. --- Any idea where it has gone, if it ever existed?? Greets, Marko [1] https://git.reviewboard.kde.org/r/123334/ ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120648: Encode the URIs which end up in DTD files
On April 3, 2015, 5:58 p.m., Andrius da Costa Ribas wrote: cmake/uriencode.cmake, line 15 https://git.reviewboard.kde.org/r/120648/diff/2/?file=320784#file320784line15 I've tried changing this to: execute_process(COMMAND perl -MURI::Escape -e print uri_escape_utf8(\${escaped_uri}\, \^A-Za-z0-9\\-\\._~\\/:\); (...) it works here, but I'm not sure if it'd be the proper fix. Patrick von Reth wrote: ah that whas the error, thanks I finally can compile kdoctools (atleast locally ) again Has this already been committed, or should this RR still considered to be re-opened? - Marko --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120648/#review78445 --- On Feb. 28, 2015, 11:02 p.m., Luigi Toscano wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120648/ --- (Updated Feb. 28, 2015, 11:02 p.m.) Review request for Build System, KDE Software on Mac OS X, KDE Frameworks, and kdewin. Repository: kdoctools Description --- The URI need to be encoded, because some valid characters for filenames are not valid according RFC 2396. Easy way to trigger the issue: when the path contains spaces, as it happens on MacOSX builds. See also https://git.reviewboard.kde.org/r/120649/ for the twin review on kdelibs4support. Diffs - cmake/uriencode.cmake PRE-CREATION src/CMakeLists.txt 341ecf4 Diff: https://git.reviewboard.kde.org/r/120648/diff/ Testing --- It compiles, but I can't properly test Mac and Windows scenarios Thanks, Luigi Toscano ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120648: Encode the URIs which end up in DTD files
On April 3, 2015, 5:58 p.m., Andrius da Costa Ribas wrote: cmake/uriencode.cmake, line 15 https://git.reviewboard.kde.org/r/120648/diff/2/?file=320784#file320784line15 I've tried changing this to: execute_process(COMMAND perl -MURI::Escape -e print uri_escape_utf8(\${escaped_uri}\, \^A-Za-z0-9\\-\\._~\\/:\); (...) it works here, but I'm not sure if it'd be the proper fix. Patrick von Reth wrote: ah that whas the error, thanks I finally can compile kdoctools (atleast locally ) again Marko Käning wrote: Has this already been committed, or should this RR still considered to be re-opened? Luigi Toscano wrote: No reopen, it was submitted. No, it was not committed, I'd like to understand the implication of the patch. Thanks for clarifying. BTW, thanks to Scarlett's progress on the new CI system it looks like we might soon be able to check whether this works also reliably on OSX when installing the files in space-containing paths. - Marko --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120648/#review78445 --- On Feb. 28, 2015, 11:02 p.m., Luigi Toscano wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120648/ --- (Updated Feb. 28, 2015, 11:02 p.m.) Review request for Build System, KDE Software on Mac OS X, KDE Frameworks, and kdewin. Repository: kdoctools Description --- The URI need to be encoded, because some valid characters for filenames are not valid according RFC 2396. Easy way to trigger the issue: when the path contains spaces, as it happens on MacOSX builds. See also https://git.reviewboard.kde.org/r/120649/ for the twin review on kdelibs4support. Diffs - cmake/uriencode.cmake PRE-CREATION src/CMakeLists.txt 341ecf4 Diff: https://git.reviewboard.kde.org/r/120648/diff/ Testing --- It compiles, but I can't properly test Mac and Windows scenarios Thanks, Luigi Toscano ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: qca fails to build on branch qt5 for branch groups (stable-)kf5-qt5
On 11 Apr 2015, at 01:22 , Marko Käning mk-li...@email.de wrote: I see tons of deprecation warnings for QCA as well as a real error: --- /Users/marko/WC/KDECI-builds/stable-kf5-qt5/qca/plugins/qca-ossl/qca-ossl.cpp:5719:13: warning: 'SSL_get_error' is deprecated: first deprecated in OS X 10.7 [-Wdeprecated-declarations] int x = SSL_get_error(ssl, ret); ^ /usr/include/openssl/ssl.h:1506:5: note: 'SSL_get_error' has been explicitly marked deprecated here int SSL_get_error(const SSL *s,int ret_code) DEPRECATED_IN_MAC_OS_X_VERSION_10_7_AND_LATER; ^ /Users/marko/WC/KDECI-builds/stable-kf5-qt5/qca/plugins/qca-ossl/qca-ossl.cpp:5808:33: error: use of undeclared identifier 'SSL_SESSION_get_compress_id' sessInfo.isCompressed = (0 != SSL_SESSION_get_compress_id(ssl-session)); ^ — It indeed fails for kf5-qt5 as well. ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: qca fails to build on branch qt5 for branch group stable-kf5-qt5
I see tons of deprecation warnings for QCA as well as a real error: --- /Users/marko/WC/KDECI-builds/stable-kf5-qt5/qca/plugins/qca-ossl/qca-ossl.cpp:5719:13: warning: 'SSL_get_error' is deprecated: first deprecated in OS X 10.7 [-Wdeprecated-declarations] int x = SSL_get_error(ssl, ret); ^ /usr/include/openssl/ssl.h:1506:5: note: 'SSL_get_error' has been explicitly marked deprecated here int SSL_get_error(const SSL *s,int ret_code) DEPRECATED_IN_MAC_OS_X_VERSION_10_7_AND_LATER; ^ /Users/marko/WC/KDECI-builds/stable-kf5-qt5/qca/plugins/qca-ossl/qca-ossl.cpp:5808:33: error: use of undeclared identifier 'SSL_SESSION_get_compress_id' sessInfo.isCompressed = (0 != SSL_SESSION_get_compress_id(ssl-session)); ^ --- qca.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
KDE/CI for KF5
Hi Ben, On 08 Mar 2015, at 04:05 , Ben Cooksley bcooks...@kde.org wrote: Not entirely. What we probably need is a better way for developers to communicate to people on the porting status. yes, I guess so. The best way to do this would probably be through the build metadata - if the branch has been set then we can probably assume developers think it is ready. Existence of the branch alone isn’t enough, as Christoph and I noticed. Anything being released needs to have CI really. If the tarballs end up on download.kde.org, it should be Green on CI first. OK, everything with a tarball on download.kde.org? Well, I will integrate such a check in my CI scripts. That is quite a bit of effort for little return. You can say that again. Out of interest, what was the primary blocker to getting things to build? 140 seems quite high as a number…. Yes, that’s a lot, more general failures are due to - porting to KF5 not yet done - missing Qt stuff (e.g. Qt5GStreamer, Qt5WebEngine, AccountsQt5, SignOnQt5) - QCA2 (even a problem on Linux) and OSX-specific: - some dependency on X11 - issues with the correct location of libs installed through MacPorts - missing dependencies on MacPorts - Apple’s clang not supported - linking issues I have left a few comments in my tier 5 list of projects [1] and more outdated on [2]. Anything in Extragear, Frameworks or kde/* should be covered by CI really. If it can't build, we need to fix that. OK. Everything IS a lot, too much in fact, so we need to focus! This must be helped by the project developers themselves. I think every project should cross-check against Christoph’s site [3] and verify the fairly detailed - automagically determined - porting status information. What about introducing a status change notification feature for [3]? That way the CI team could get informed once a project newly arrives or somehow updates (by e.g. removing kdelibs4support) on KF5? I think this would be a real help for a more effective KDE/CI. Greets, Marko [1] http://quickgit.kde.org/?p=clones%2Fwebsites%2Fbuild-kde-org%2Fkaning%2Fmp-osx-ci.gita=blobh=d582668b5ddf434948b39257bf4db673b9f69addhb=daa9eafdc6186a5a60cc46ac3dd9e23e5f76611df=tier5.fw [2] https://trac.macports.org/wiki/KDEProblems/KDEMacPortsCI/Status/ProjectsBeyondKF5 [3] http://developer.kde.org/~cfeck/portingstatus.html ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KDE Frameworks 5.8
On 07 Mar 2015, at 19:32 , šumski hrvoje.sen...@gmail.com wrote: A few things regarding kpeople: It requires Qt 5.3, rest of KF5 5.2; There are quite a few frameworks which require 5.4 even! signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
KDE/CI for KF5 WAS OSX/CI: libkcddb and ark fail to build for KF5
Hi Jeremy AND folks, On 07 Mar 2015, at 02:14 , Jeremy Whiting jpwhit...@kde.org wrote: I appreciate the efforts you are making :) Thanks. :) Just wondering whether I am wasting my time because of my “perfectionism here on OSX/CI. E.g. libkcddb I perhaps shouldn’t have touched at all for now, as it hasn’t been committed to since last autumn… Well, for ark it’s different - of course! But still, e.g. on my dependencies-RR [1] I have been working for 2 weeks now and while doing so I more and more began to wonder whether it makes sense for all the new projects I am trying to get into CI... I have the feeling that my work (inspired by Christoph Feck’s post on KDE-DEVEL [2] and his porting status page [3]) runs a little too uncoordinated - which gives me only extra head aches, since I have to iterate through all these projects _manually_ while knowing far too little of their real status and objectives. This lead to the result that I - could effectively only add about _30_ new projects - which means that there are now about 200 projects successfully built on OSX/CI [4], - while having had to go through *170* potential projects from [3] step by step. For all of these I needed to figure out how to build them on OSX/CI and how to fix their dependency metadata. 140 projects are thus disabled on OSX/CI for now. I doubt that this can be considered a very efficient workflow! ;-) I think we need to have another CI IRC meeting sometime soon, where we could discuss - besides the change to the new KDE/CI system - also which projects are actually worth considering (not only on OSX but) on Linux/CI in general! What do you think? Greets, Marko [1] https://git.reviewboard.kde.org/r/122672/ [2] http://lists.kde.org/?l=kde-develm=142434697530863w=2 [3] http://developer.kde.org/~cfeck/portingstatus.html [4] https://trac.macports.org/wiki/KDEProblems/KDEMacPortsCI/Status ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
Dear Michael, On 04 Mar 2015, at 02:00 , Michael Pyne mp...@kde.org wrote: There's a reason I'd mentioned kdesrc-build's current behavior in my reply to that RR. :) yep, I figured that. And thus I decided to come back to this issue - targeting a wider audience. :) I personally would retain empty entries. OK, I see, it makes sense to revert all these removals then, just to make sure that the current state of kdesrc-build behaves as it was supposed to do with these explicit defines… But that would be a significant behavior change, especially for lesser-used modules (e.g. in playground/) that don't necessarily receive CI coverage, but which users and developers may still want to build via kdesrc-build. Yes, I see that point. I guess the majority of projects currently does NOT have CI… So, imposing the CI’s way of dealing with dependencies would add a lot of extra work on all devs… I think the real solution (so that we don't need empty branch-group hacks) would come from finally implementing the proposal Ben and I had made back in August 2014 (currently just an email thread in kde-frameworks-devel https://mail.kde.org/pipermail/kde-frameworks-devel/2014-August/018391.html). Yes, I remember your post from last August, but unfortunately there was no (visual) feedback on the list, nor is there any feedback here on K-F-D up to now - which is no good news, I suppose. I ran out of time to do effectively any development for some months after that, so as far as I know there's been no progress. But that's the direction we *intend* to head... now would be a good time if you want to review the proposal to see if it would help or hurt your efforts. Right, I do believe it is time to come up with a decision in this respect! In the light of Ben’s and Scarlett’s work on the new (multi-platform) KDE/CI system it would be good to also tackle dependency definitions for kdesrc-build. Scarlett has already acted in this regard, so it would be good if we could make that more consistent for kdesrc-build as well. Yet, I am afraid the subject of this thread isn’t attracting too many readers. Probably you should simply repost your proposal from last August and ask for any active discussion. I believe CI is a critical component for KDE’s further development and thus shouldn’t be treated with too much neglect. ;-) Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
Hi Michael, On 03 Mar 2015, at 03:30 , Michael Pyne mp...@kde.org wrote: On Mon, March 2, 2015 20:15:20 Marko Käning wrote: So, having those ktp-* entries in dependency-data-stable-kf5-qt5 shouldn’t do any harm. But I don’t know how kdesrc-build will handle this, though... For missing entries kdesrc-build will infer master as the git branch to use. It would be very confusing for a user to ask to build a module and still have kdesrc-build ignore it because it doesn't have an entry in kde-build-metadata. With that said, kdesrc-build *will* ignore modules that have a defined branch of (i.e. empty) in logical-module-structure, so if a module simply should not be built for a given branch-group my recommendation would be to define the branch-group after all but set it to an empty value. E.g. kde/kdenetwork/ktp*: { stable-qt4: kde-telepathy-0.9, latest-qt4: kde-telepathy-0.9, kf5-qt5: master, stable-kf5-qt5: }, I believe that Scarlett's new CI supports this as well, and the current Jenkins CI also supports this. Scarlett’s CI also supports to treat *undefined* entries as _set to empty_, just like my OSX/CI does. So, in the light of your remarks the question is, whether all the removed empty definitions in my RR [1] should actually be left the way they are!?!? Greets, Marko [1] https://git.reviewboard.kde.org/r/122672/ ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
Hi, I think CI still needs something like this: --- $ git diff diff --git a/dependency-data-kf5-qt5 b/dependency-data-kf5-qt5 index 3731ec6..6005971 100644 --- a/dependency-data-kf5-qt5 +++ b/dependency-data-kf5-qt5 @@ -241,6 +241,11 @@ frameworks/plasma-framework: frameworks/kwindowsystem frameworks/plasma-framework: frameworks/kxmlgui frameworks/plasma-framework: frameworks/ktexteditor frameworks/kxmlrpcclient: frameworks/kio +frameworks/kpeople: frameworks/kcoreaddons +frameworks/kpeople: frameworks/kwidgetsaddons +frameworks/kpeople: frameworks/kservice +frameworks/kpeople: frameworks/ki18n +frameworks/kpeople: frameworks/kitemviews # Frameworks, tier4 frameworks/frameworkintegration: frameworks/ki18n @@ -1056,12 +1061,5 @@ kde/kdegames/palapeli: kde/kdegames/libkdegames kde/kdegames/picmi: kde/kdegames/libkdegames extragear/games/knights: kde/kdegames/libkdegames -# Playground Libs -playground/network/kpeople: frameworks/kcoreaddons -playground/network/kpeople: frameworks/kwidgetsaddons -playground/network/kpeople: frameworks/kservice -playground/network/kpeople: frameworks/ki18n -playground/network/kpeople: frameworks/kitemviews - # A standalone application, no need to install KF5 extragear/pim/trojita: -frameworks/kf5umbrella diff --git a/dependency-data-stable-kf5-qt5 b/dependency-data-stable-kf5-qt5 index 912255e..d5ed49b 100644 --- a/dependency-data-stable-kf5-qt5 +++ b/dependency-data-stable-kf5-qt5 @@ -236,6 +236,11 @@ frameworks/plasma-framework: frameworks/kwidgetsaddons frameworks/plasma-framework: frameworks/kwindowsystem frameworks/plasma-framework: frameworks/kxmlgui frameworks/plasma-framework: frameworks/ktexteditor +frameworks/kpeople: frameworks/kcoreaddons +frameworks/kpeople: frameworks/kwidgetsaddons +frameworks/kpeople: frameworks/kservice +frameworks/kpeople: frameworks/ki18n +frameworks/kpeople: frameworks/kitemviews # Frameworks, tier4 frameworks/frameworkintegration: frameworks/ki18n --- Greets, Marko___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
Hi, On 02 Mar 2015, at 19:14 , Luigi Toscano luigi.tosc...@tiscali.it wrote: This should really be solved somehow in a more clear way, if I'm not misunderstanding all the process: there is no stable version of ktp right now for kf5. yes, I guess so. If there is no stable version for it, it shouldn’t appear in this file! Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
We need to remember that the stable-kf5-qt5 branch group needs to be kept in sync whenever updating kf5-qt5… I committed the still missing info for it in [1]. Greets, Marko [1] http://commits.kde.org/kde-build-metadata/1aa3609df9b3b513c4fd555d38cb3d047f4e3c03 ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KPeople part of KDE Frameworks
Well, Luigi, this is what the logical deps say: --- extragear/network/telepathy/*: { stable-qt4: kde-telepathy-0.9, latest-qt4: kde-telepathy-0.9 }, kde/kdenetwork/ktp*: { stable-qt4: kde-telepathy-0.9, latest-qt4: kde-telepathy-0.9, kf5-qt5: master }, --- which means that kde-telepathy-0.9” is set as the latest stable for Qt4 only. As the port group “stable-kf5-qt5” isn’t set to a value, it is not considered in OSX/CI and Scarlett’s new version of Linux/CI. So, having those ktp-* entries in dependency-data-stable-kf5-qt5 shouldn’t do any harm. But I don’t know how kdesrc-build will handle this, though... Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: purpose fails to build on branch master
Hi Aleix, after adding a few dependencies: --- $ git diff diff --git a/dependency-data-kf5-qt5 b/dependency-data-kf5-qt5 index 7cf844f..2a303d4 100644 --- a/dependency-data-kf5-qt5 +++ b/dependency-data-kf5-qt5 @@ -1002,6 +1002,12 @@ playground/base/kio-mtp: frameworks/ki18n playground/base/kio-mtp: frameworks/solid playground/base/kio-mtp: frameworks/kio +# Playground Libs +playground/libs/purpose: frameworks/kcoreaddons +playground/libs/purpose: frameworks/kconfig +playground/libs/purpose: frameworks/ki18n +playground/libs/purpose: frameworks/kdeclarative + # KAccounts kdereview/kaccounts-integration: frameworks/kcmutils kdereview/kaccounts-integration: frameworks/kio diff --git a/logical-module-structure b/logical-module-structure index c7f929e..284c54a 100644 --- a/logical-module-structure +++ b/logical-module-structure @@ -738,6 +738,9 @@ playground/libs/kpackage : { kf5-qt5: master }, +playground/libs/purpose : { +kf5-qt5: master +}, playground/base/qtcurve : { latest-qt4: master, kf5-qt5: “master” --- I was able to start building purpose on OSX/CI... ...but then it failed later on: --- /Users/marko/WC/KDECI-builds/kf5-qt5/purpose/src/plugins/ktp-sendfile/ktpsendfileplugin.cpp:65:28: error: no matching constructor for initialization of 'const QJsonObject ' Q_EMIT output( {{ QStringLiteral(url), {} }}); ^~~ /Users/marko/WC/KDECI-builds/kf5-qt5/purpose/src/plugins/dummy/dummyplugin.cpp:48:13: error: no matching function for call to 'singleShot' QTimer::singleShot(100, this, [this](){ setPercent(10); }); ^~ --- Please find the whole build log attached. Greets, Marko P.S.: I was trying to build kamoso. Stop me, please, if this still doesn’t make sense (on OSX)!!! purpose.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121390: make Qt5 theme build on Linux again
engine present doesn't mean that indidual KDE applications or plugins can no longer decide to support only a subset to be set at build time. *) No issue either with Unix but not OS X - that's what I came up with for lack of something better. Turns out Yichao has his own alternative to HAVE_X11, I'll see if I can make do with that. *) or else I'll start making a ruckus to have kwin and more Plasma goodies on OS X!! ;) Martin Gräßlin wrote: Yes, it's not about compile time checks, it's about introducing runtime checks as Thomas and I wrote ;-) René J.V. Bertin wrote: Actually, Thomas wrote The compile time checks have no implication on the runtime. Surely a typo, but those can have devastating consequences around code ;) René J.V. Bertin wrote: (published too fast again) Actually, that blog post of yours also starts out by talking exclusively about compile-time checks for about 2/3 of its length. It's only after the screenshot that it becomes clear you actually use the compile-time check to include a runtime-check or not. A casual reader might be tempted to stop reading early, thinking he got the message ... And I can't stop thinking something that has been stamped into me: ifs are expensive. Guess that shows my age ... Thomas Lübking wrote: That's not a typo. Meaning distorting partial quote. I wrote: The compile time checks have no implication on the runtime *environment*. Ifs are expensive might be stamped into your mind and/or true, but they're completely inavoidable in this context. Just that X11 was available at runtime does NOT (no more w/ Qt5) mean that it's available at runtime. = Keep the branching out of hot loops (as always) ;-) René J.V. Bertin wrote: yes, I know I didn't copy the last word of your statement. That doesn't change the fact that your 2nd word was *compile* instead of *run*, in a context where you (at least) seemed to be saying that I apparently claimed that those (= compile time) checks had an impact on runtime performance. Anyway, yes, I understood perfectly well that X11 might not be available at runtime while it was when compiling, and that an application trying to do X11 calls will exit with an error when trying to connect to an inexisting X11 server. (Or crash if X11 was actually uninstalled ... but it would take other runtime checks to protect against that, and frankly that'd just be crazy.) Ifs are expensive might be stamped into your mind and/or true, but they're completely inavoidable in this context. Not true, see my remarks about using function pointers above. Not that that would be particularly clever and less expensive when X11 is the only platform that provides a certain functionality ... :) (I do seem to recall that using function pointers instead of normal functions was hardly more expensive on x86) Yichao Yu wrote: Sorry somehow my filter missed this review request and I've just seen it today... To answer Martin's concern, I totally agree and it's in my mind the first time I added X11 support back to the Qt5 version. The X11 related stuff in `libqtcurve-utils` have also always had that check. All X11 related functions are guarded by both a compile time check (e.g. if libxcb/X11 is not found or somehow the user simply don't want to link to them...) and a runtime check (i.e. the X11 related functions are no-ops if X11 is not initialized first at runtime). Now (AFAIK) the compile failure on OSX seems to be related to some sturture name conflict (or whatever it is that causes a forward declaration of `Display` not working...). The real issue is already addressed in another review request and it is not necessary to disable calls to X11 related functions (which might be no-ops) on OSX anymore. In any case, the issue related to this request should already be resolved now and the status is also monitored on build.kde.org (and AFAIK both Qt4 and Qt5 versions build successfully now). I think this review request can be discarded. Marko Käning wrote: Just for the record, QtCurve currently fails to build on OSX/CI: /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/qt5/config/qtcurveconfig.cpp:1085: Warning: Macro argument mismatch. In file included from /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/lib/utils/dirs.cpp:22: /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/lib/utils/dirs.h:68:1: error: implicit instantiation of undefined template 'std::__1::basic_stringchar, std::__1::char_traitschar, std::__1:: allocatorchar ' getConfFile(const std::string file) ^ Shall I send the full build log of this failure to you via PM? For completeness: I haven't THIS RR applied to my OSX/CI system as of now. SHOULD I, perhaps??? - Marko
Re: Review Request 121390: make Qt5 theme build on Linux again
On Dec. 8, 2014, 3:07 p.m., Martin Gräßlin wrote: this is wrong - please have a look at various frameworks on how to do it properly. In the end it should be: #if HAVE_X11 // x11 specific stuff #endif obviously it also needs a runtime check: if (QX11Info::isPlatformX11()) as we no longer should assume that X11 is the only platform on Unix(non OSX). René J.V. Bertin wrote: I found a reference to HAVE_X11 online, but that token is not defined. Note also that the Qt5 theme is supposed to build without KF5 too, for pure Qt5 applications, so this kind of token should rather be provided by the Qt cmake files rather than KDE's. I'll leave the runtime checks to the QtCurve devs, as they obviously should be made in lots of locations and it's their call. I myself don't see the point in doing a runtime check whether a platform is X11, though. It's known at build time if a platform is X11. Doing a runtime check before each theming action if `Q11Info::isX11Active()` (or some such call) seems to be an expensive concession to a rather improbable set-up. If distros really decide to give the user a choice between X11 and Wayland at login ... let them figure out how to streamline that first, and then add runtime checks for the active graphics backend where really needed... (In fact, I myself would avoid anything tied to the display layer as much as possible; the fact I had to go back in a few months after the previous porting effort goes to show how easy it is to break builds on other platforms with that kind of functionality - as if my own error wasn't enough already.) Martin Gräßlin wrote: HAVE_X11 is neither defined by Qt5 nor by KF5. It needs to be set manually depending on whether the source is built with or without X11 support. Concerning the runtime check: kwrite -platform wayland and your app is going to crash like hell if there is no runtime checks. René J.V. Bertin wrote: ``` neon5-exec /opt/project-neon5/bin/kwrite -platform wayland This application failed to start because it could not find or load the Qt platform plugin wayland. Available platform plugins are: linuxfb, minimal, offscreen, xcb. Reinstalling the application may fix this problem. Abort (core dumped) ``` Right, so with runtime checks it doesn't crash, it just self-destructs. Very useful difference :) Of course an application will crash if it tries to use an unavailable displaying method, or if the linker tries to load shared libraries that aren't present. In fact, with X11 it might actually exit nicely with a message about a display rather than crash. This just underlines my point: the only use for runtime checks in this context if is the same binaries are supposed to work with multiple displaying backends (or platform plugins). Whether QtCurve ought to support that is a call for its developers to make, like I said above. The only way to do that properly without (too much) overhead is to do the check at initialisation time rather than preceding each backend-specific call, i.e. use functionpointers or some OO/C++ alternative. I don't know how preferable Wayland is to X11; without that I see only an interest for people working on Wayland development under X11 for this kind of runtime switch support. To put this into context: I've often thought how it'd be nice if Qt-mac would be able to use X11, but I've always dismissed the possibility that that might be a runtime switch, exactly because it would introduce too much overhead and/or complexity for a feature that'd be used only rarely. Thomas Lübking wrote: Right, so with runtime checks it doesn't crash, it just self-destructs. You're missing the point entirely. The compile time checks have no implication on the runtime environment. Of course you cannot use the wayland platform plugin if it's not available, but you *can* compile Qt/KDE w/ X11 and wayland present - but making X11 calls when running on the wayland PP will crash the application - thus you must check whether you're operating on X11/xdg at runtime. Also testing for UNIX but not OSX to make X11 calls is plain wrong. Could be framebuffer console or wayland and no X11 installed at all. Martin Gräßlin wrote: for more information please see my blog post: http://blog.martin-graesslin.com/blog/2014/02/running-frameworks-powered-applications-on-wayland/ Btw. the QtWayland PPA will be available starting with Qt 5.4 - a version I'm already using. René J.V. Bertin wrote: @Thomas: we're not talking about compile time checks. Those evidently don't have any implication on the runtime environment (if done correctly :)). I guess my point is simply that the fact that you can (= it's possible to) compile Qt/KDE with every conceivable display/rendering
Re: Review Request 121390: make Qt5 theme build on Linux again
engine present doesn't mean that indidual KDE applications or plugins can no longer decide to support only a subset to be set at build time. *) No issue either with Unix but not OS X - that's what I came up with for lack of something better. Turns out Yichao has his own alternative to HAVE_X11, I'll see if I can make do with that. *) or else I'll start making a ruckus to have kwin and more Plasma goodies on OS X!! ;) Martin Gräßlin wrote: Yes, it's not about compile time checks, it's about introducing runtime checks as Thomas and I wrote ;-) René J.V. Bertin wrote: Actually, Thomas wrote The compile time checks have no implication on the runtime. Surely a typo, but those can have devastating consequences around code ;) René J.V. Bertin wrote: (published too fast again) Actually, that blog post of yours also starts out by talking exclusively about compile-time checks for about 2/3 of its length. It's only after the screenshot that it becomes clear you actually use the compile-time check to include a runtime-check or not. A casual reader might be tempted to stop reading early, thinking he got the message ... And I can't stop thinking something that has been stamped into me: ifs are expensive. Guess that shows my age ... Thomas Lübking wrote: That's not a typo. Meaning distorting partial quote. I wrote: The compile time checks have no implication on the runtime *environment*. Ifs are expensive might be stamped into your mind and/or true, but they're completely inavoidable in this context. Just that X11 was available at runtime does NOT (no more w/ Qt5) mean that it's available at runtime. = Keep the branching out of hot loops (as always) ;-) René J.V. Bertin wrote: yes, I know I didn't copy the last word of your statement. That doesn't change the fact that your 2nd word was *compile* instead of *run*, in a context where you (at least) seemed to be saying that I apparently claimed that those (= compile time) checks had an impact on runtime performance. Anyway, yes, I understood perfectly well that X11 might not be available at runtime while it was when compiling, and that an application trying to do X11 calls will exit with an error when trying to connect to an inexisting X11 server. (Or crash if X11 was actually uninstalled ... but it would take other runtime checks to protect against that, and frankly that'd just be crazy.) Ifs are expensive might be stamped into your mind and/or true, but they're completely inavoidable in this context. Not true, see my remarks about using function pointers above. Not that that would be particularly clever and less expensive when X11 is the only platform that provides a certain functionality ... :) (I do seem to recall that using function pointers instead of normal functions was hardly more expensive on x86) Yichao Yu wrote: Sorry somehow my filter missed this review request and I've just seen it today... To answer Martin's concern, I totally agree and it's in my mind the first time I added X11 support back to the Qt5 version. The X11 related stuff in `libqtcurve-utils` have also always had that check. All X11 related functions are guarded by both a compile time check (e.g. if libxcb/X11 is not found or somehow the user simply don't want to link to them...) and a runtime check (i.e. the X11 related functions are no-ops if X11 is not initialized first at runtime). Now (AFAIK) the compile failure on OSX seems to be related to some sturture name conflict (or whatever it is that causes a forward declaration of `Display` not working...). The real issue is already addressed in another review request and it is not necessary to disable calls to X11 related functions (which might be no-ops) on OSX anymore. In any case, the issue related to this request should already be resolved now and the status is also monitored on build.kde.org (and AFAIK both Qt4 and Qt5 versions build successfully now). I think this review request can be discarded. Marko Käning wrote: Just for the record, QtCurve currently fails to build on OSX/CI: /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/qt5/config/qtcurveconfig.cpp:1085: Warning: Macro argument mismatch. In file included from /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/lib/utils/dirs.cpp:22: /Users/marko/WC/KDECI-builds/kf5-qt5/qtcurve/lib/utils/dirs.h:68:1: error: implicit instantiation of undefined template 'std::__1::basic_stringchar, std::__1::char_traitschar, std::__1:: allocatorchar ' getConfFile(const std::string file) ^ Shall I send the full build log of this failure to you via PM? Marko Käning wrote: For completeness: I haven't THIS RR applied to my OSX/CI system as of now
Re: [kio] /: Look for kdesu in the correct location
Hi Marco, I’ve seen your below commit for kio and wondered whether something similar regarding CMAKE_INSTALL_FULL_LIBEXECDIR could be done for kconfig, since currently I need this OSX-specific configure option: --- $ cat config/build/darwin-mavericks.cfg [DEFAULT] configureExtraArgs=-DCMAKE_INSTALL_BUNDLEDIR=lib/libexec/kf5 --- to make sure that kconfig can actually find kconfig_compiler.app which otherwise gets installed in an OSX-location [1]. Surely this application doesn’t need to be exposed to a user, thus the location in $PREFIX/lib/libexec/kf5 is ok. But in case cmake does install it - like in [1] below the path /Applications/KDE - kconfig and kservice [2] should also be able to find kconfig_compiler's application package. Greets, Marko [1] https://paste.kde.org/pdubpj7nm [2] https://mail.kde.org/pipermail/kde-frameworks-devel/2014-May/016059.html On 09 Feb 2015, at 18:41 , Marco Martin notm...@gmail.com wrote: Git commit 5d70bd5de594ff40a009c0e783f9a0a7eed2f14e by Marco Martin. Committed on 09/02/2015 at 17:39. Pushed by mart into branch 'master'. Look for kdesu in the correct location Look for kdesu under CMAKE_INSTALL_FULL_LIBEXECDIR directory patch by Maarten De Meyer REVIEW:120185 CCMAIL:de.meyer.maar...@gmail.com M +2-2autotests/krununittest.cpp M +11 -1src/core/desktopexecparser.cpp http://commits.kde.org/kio/5d70bd5de594ff40a009c0e783f9a0a7eed2f14e diff --git a/autotests/krununittest.cpp b/autotests/krununittest.cpp index 73a3932..03cbde8 100644 --- a/autotests/krununittest.cpp +++ b/autotests/krununittest.cpp @@ -154,8 +154,8 @@ void KRunUnitTest::testProcessDesktopExec() int pt = ex + te * 2 + su * 4; QString exe; if (pt == 4 || pt == 5) { -exe = QStandardPaths::findExecutable(kdesu); -if (exe.isEmpty()) { +exe = QFile::decodeName(CMAKE_INSTALL_FULL_LIBEXECDIR_KF5 /kdesu); +if (!QFile::exists(exe)) { qWarning() kdesu not found, skipping test; continue; } diff --git a/src/core/desktopexecparser.cpp b/src/core/desktopexecparser.cpp index e20f046..7bb29ba 100644 --- a/src/core/desktopexecparser.cpp +++ b/src/core/desktopexecparser.cpp @@ -393,7 +393,17 @@ QStringList KIO::DesktopExecParser::resultingArguments() const if (d-service.terminal()) { result su; } else { -result QStandardPaths::findExecutable(kdesu) -u; +QString kdesu = QFile::decodeName(CMAKE_INSTALL_FULL_LIBEXECDIR_KF5 /kdesu); +if (!QFile::exists(kdesu)) { +kdesu = QStandardPaths::findExecutable(kdesu); +} +if (!QFile::exists(kdesu)) { +// Insert kdesu as string so we show a nice warning: 'Could not launch kdesu' +result QStringLiteral(kdesu); +return result; +} else { +result kdesu -u; +} } result d-service.username() -c; ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build for branch-group stable-kf5-qt5
Hi Albert, On 15 Feb 2015, at 23:12 , Albert Astals Cid aa...@kde.org wrote: El Divendres, 13 de febrer de 2015, a les 08:49:10, Marko Käning va escriure: Hi Albert, I just realised that kde-baseapps fails on OSX/CI when being build for branch group stable-kf5-qt5: Because there's no such thing as kde-baseapps stable-kf5-qt5, your scripts assume there is, but that's in my opinion your scripts assuming too much on what kde-build-metadata says. yes, you’re absolutely right. OTOH Ben agrees with you and since he created the scritpts, fine, let's assume one can try to build branches of repos that don't make sense, so please fix things to suit your way, but make sure you're not breaking things that we have actually released. He has set this by now: --- 04833a43 (Ben Cooksley 2015-02-13 22:58:54 +1300 624) stable-kf5-qt5: “ --- and OSX/CI is happy again - and NOT building anything which doesn’t make sense. So, after all, nothing to worry about! :) More important was that two frameworks (attica and plasma-framework) had nothing set for the stable-kf5-qt5 branch! Obviously Linux/CI assumes ‘master' if nothing is set, but OSX/CI doesn’t and thus stumbled over these two. This is now fixed in [1]. Greets, Marko [1] http://quickgit.kde.org/?p=kde-build-metadata.gita=blobdiffh=ecf795a6ae171b1804553386d574814c531a4c10hp=772dac0e613a754baa901e91954ff9d88ef0hb=7918a6119acbbce72d20052ef1ad14d75353a93ff=logical-module-structure ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build for branch-group stable-kf5-qt5
OK, I understood only now, that kde-baseapps [1] and likely also the other two projects are KDE4-only. There is no EMPTY entry for “stable-kd5-qt5” branch group for these three projects in our logical dependencies, which should be the reason for this. Should I add those? (I suspect also that I’ll stumble over more than these 3…) P.S.: Just now I ran into ark and pairs failing as well! ;-) ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build for branch-group stable-kf5-qt5
Hi, On 13 Feb 2015, at 10:15 , Kevin Funk kf...@kde.org wrote: Indeed, the kde/* rule in 'logical-module-structure' is wrong already: yep, that’s what I wanted to express. =) Ben has fixed that in the meantime [1] and then changed it later again [2]. So, now one would have to introduce for all the below projects a separate dependency block to ensure that “stable-kf5-qt5” is set to empty: == FAILURE SUMMARY: libkdcraw kqtquickcharts pairs ark cantor kmplot kolourpaint kmahjongg granatier killbots ktuberling knavalbattle lskat palapeli kreversi ksirk kturtle ksnakeduel knetwalk ksudoku kubrick ksquares kspaceduel katomic kbreakout kfourinline kgoldrunner kiriki kjumpingcube klickety kolf kollision konquest kpat kigo == It seems to me, though, that at least something like this --- kde/kdegames/*: { stable-qt4: Applications/14.12, latest-qt4: Applications/14.12, kf5-qt5: frameworks, stable-kf5-qt5: } --- might be more appropriate as a default setting, no!? Greets, Marko [1] http://quickgit.kde.org/?p=kde-build-metadata.gita=commitdiffh=f2332e41a47184f43946675d01b05d3ece74d7ec [2] http://quickgit.kde.org/?p=kde-build-metadata.gita=commitdiffh=04833a43f38bc5d316102b6fd675d0d177542e19 signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kde-baseapps fails to build for branch-group stable-kf5-qt5
Hi Albert, I just realised that kde-baseapps fails on OSX/CI when being build for branch group stable-kf5-qt5: -- Group: stable-kf5-qt5 Project : kde/applications/kde-baseapps Branch : Applications/14.12 Linux-CI : UNSTABLE OSX/CI : FAILURE Prep... Build... : BUILD FAILED == This I find in the build log: --- -- The C compiler identification is AppleClang 6.0.0.656 -- The CXX compiler identification is AppleClang 6.0.0.656 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done CMake Error at /opt/kde/install/darwin/mavericks/clang/shared/general/cmake/share/cmake-3.1/Modules/FindKDE4.cmake:71 (message): ERROR: Could not find KDE4 kde4-config Call Stack (most recent call first): CMakeLists.txt:10 (find_package) CMake Warning (dev) in CMakeLists.txt: No cmake_minimum_required command is present. A line of code such as cmake_minimum_required(VERSION 3.1) should be added at the top of the file. The version specified may be lower if you wish to support older CMake versions for this project. For more information run cmake --help-policy CMP. This warning is for project developers. Use -Wno-dev to suppress it. -- Configuring incomplete, errors occurred! See also /Users/marko/WC/KDECI-builds/stable-kf5-qt5/kde-baseapps/build/CMakeFiles/CMakeOutput.log. KDE Continuous Integration Build == Building Project: kde-baseapps - Branch Applications/14.12 --- Any idea what goes wrong? Regards, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build for branch-group stable-kf5-qt5
I just realise that the same happens for these two projects: -- Group: stable-kf5-qt5 Project : kde/kdegraphics/libs/libkdcraw Branch : Applications/14.12 Linux-CI : SUCCESS OSX/CI : FAILURE Prep... Build... : BUILD FAILED == -- Group: stable-kf5-qt5 Project : kde/kdeedu/kqtquickcharts Branch : Applications/14.12 Linux-CI : SUCCESS OSX/CI : FAILURE Prep... Build... : BUILD FAILED == while many other build just fine, like the kde-edu projects! ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build on branch frameworks
Hi Albert, On 10 Feb 2015, at 23:17 , Jeremy Whiting jpwhit...@kde.org wrote: Exactly, whoever added that code either needs to find a Qt 5.3 way to do the same, or bump the requirement. who is in charge of kde-baseapps? I’d like to get this going on OSX with Qt 5.3.2 again, or at least see the project set its requirements correctly… Greets, Marko___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build on branch frameworks
Hi Jeremy, On 10 Feb 2015, at 23:10 , Jeremy Whiting jpwhit...@kde.org wrote: Seems that's from Qt 5.4 This enum was introduced or modified in Qt 5.4. in the QUrl::UserInputResolutionOptions documentation. Shouldn’t the project then require Qt 5.4, which I’d like to see avoided for now… Greets, Marko___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kde-baseapps fails to build on branch frameworks
To make kde-baseapps build again on OSX/CI I have removed poppler from its dependencies in config/base/kf5-qt5: --- +# KDE/Applications +kde/applications/kde-baseapps: -kde/kdelibs/baloo +kde/applications/kde-baseapps: -kde/kdelibs/baloo-widgets +kde/applications/kde-baseapps: -general/poppler --- Yet, I am running in an unrelated problem, as it seems: --- [ 75%] Building CXX object dolphin/src/CMakeFiles/kcm_dolphinviewmodes.dir/kcm_dolphinviewmodes_automoc.cpp.o /Users/marko/WC/KDECI-builds/kf5-qt5/kde-baseapps/dolphin/src/main.cpp:108:68: error: no member named 'AssumeLocalFile' in 'QUrl' const QUrl url = QUrl::fromUserInput(str, QString(), QUrl::AssumeLocalFile); ~~^ Scanning dependencies of target kitemlistselectionmanagertest 1 error generated. make[2]: *** [dolphin/src/CMakeFiles/kdeinit_dolphin.dir/main.cpp.o] Error 1 make[1]: *** [dolphin/src/CMakeFiles/kdeinit_dolphin.dir/all] Error 2 --- Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 122493: Use qRound rather than C++11 std::lround.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/#review75724 --- Yes, this lets it build on OSX/CI. Thanks, Jeremy! - Marko Käning On Feb. 9, 2015, 4:31 p.m., Jeremy Whiting wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/ --- (Updated Feb. 9, 2015, 4:31 p.m.) Review request for KDE Frameworks and Marco Martin. Repository: kdeclarative Description --- Use qRound rather than C++11 std::lround. Removes dependency on C++11. Diffs - src/qmlcontrols/kquickcontrolsaddons/plotter.cpp 67ce63a943234b167165b0f3986f974bba5ff0cf Diff: https://git.reviewboard.kde.org/r/122493/diff/ Testing --- kdeclarative is able to build on OS X 10.7 with the built in XCode compiler and standard library. Thanks, Jeremy Whiting ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 122493: Use qRound rather than C++11 std::lround.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/#review75725 --- Ship it! Ship It! - Marko Käning On Feb. 9, 2015, 4:31 p.m., Jeremy Whiting wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/ --- (Updated Feb. 9, 2015, 4:31 p.m.) Review request for KDE Frameworks and Marco Martin. Repository: kdeclarative Description --- Use qRound rather than C++11 std::lround. Removes dependency on C++11. Diffs - src/qmlcontrols/kquickcontrolsaddons/plotter.cpp 67ce63a943234b167165b0f3986f974bba5ff0cf Diff: https://git.reviewboard.kde.org/r/122493/diff/ Testing --- kdeclarative is able to build on OS X 10.7 with the built in XCode compiler and standard library. Thanks, Jeremy Whiting ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: ark fails to build on branch frameworks
Hi Raphael, On 14 Jan 2015, at 22:11 , Raphael Kubo da Costa rak...@freebsd.org wrote: Is the full build log available somewhere, preferably with `make VERBOSE=1'? This really smells like OS X having an older ( 3.0) version of libarchive in its base system, a more recent version being installed by `port' (found by CMake) and the older version being picked up for linking. yes, that’s correct. The system’s libarchive is being found. Will send details as PM. Thanks, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: ark fails to build on branch frameworks
I hadn’t built ark for quite a while (since I had activated it at some stage due to [1,2]) and only now I realise that it fails here on OSX: --- [ 69%] [ 69%] Built target kerfuffle_clirar Building CXX object plugins/clirarplugin/autotests/CMakeFiles/clirartest.dir/clirartest_automoc.cpp.o Linking CXX shared module kerfuffle_libarchive.so Scanning dependencies of target kerfuffle_clilha Scanning dependencies of target kerfuffle_karchive Undefined symbols for architecture x86_64: _archive_filter_code, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_filter_name, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_read_free, referenced from: LibArchiveInterface::ArchiveReadCustomDeleter::cleanup(archive*) in libarchivehandler.cpp.o _archive_read_support_filter_all, referenced from: LibArchiveInterface::list() in libarchivehandler.cpp.o LibArchiveInterface::copyFiles(QListQVariant const, QString const, QHashQString, QVariant) in libarchivehandler.cpp.o LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_add_filter_bzip2, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_add_filter_gzip, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_add_filter_lzma, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_add_filter_none, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_add_filter_xz, referenced from: LibArchiveInterface::addFiles(QStringList const, QHashQString, QVariant const) in libarchivehandler.cpp.o LibArchiveInterface::deleteFiles(QListQVariant const) in libarchivehandler.cpp.o _archive_write_free, referenced from: LibArchiveInterface::ArchiveWriteCustomDeleter::cleanup(archive*) in libarchivehandler.cpp.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[2]: *** [plugins/libarchive/kerfuffle_libarchive.so] Error 1 make[1]: *** [plugins/libarchive/CMakeFiles/kerfuffle_libarchive.dir/all] Error 2 --- I have this version of libarchive installed: --- $ port installed libarchive The following ports are currently installed: libarchive @3.1.2_0 (active) $ port livecheck libarchive --- which is up-to-date. What’s going on? Regards, Marko [1] https://bugreports.qt-project.org/browse/QTBUG-42605 [2] https://bugreports.qt-project.org/browse/QTBUG-42870 ark.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kactivities fails to build on branch master
Hi Ivan, On 11 Jan 2015, at 22:02 , Ivan Čukić ivan.cu...@kde.org wrote: p.s. You can try building with -DKACTIVITIES_ENABLE_EXCEPTIONS thanks very much!!! Introducing a new config build file on the CI system solved this for me on OSX! Regards, Marko --- $ git diff fd2d2caa5f640d4f416c7cc32e916eaa518b857f diff --git a/config/build/kactivities/darwin-mavericks.cfg b/config/build/kactivities/darwin-mavericks.cfg new file mode 100644 index 000..5512f48 --- /dev/null +++ b/config/build/kactivities/darwin-mavericks.cfg @@ -0,0 +1,2 @@ +[DEFAULT] +configureExtraArgs=-DKACTIVITIES_ENABLE_EXCEPTIONS:BOOL=TRUE ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kdeclarative fails to build on branch master
Hi, kdeclarative currently fails to build on OSX/CI because I haven’t installed port lib epoxy yet. The question though is: is it really needed on OSX, as X11 isn’t really used with a native OSX/Qt install?? Regards, Marko -- The following REQUIRED packages have not been found: * epoxy , libepoxy , http://github.com/anholt/libepoxy OpenGL dispatch library CMake Error at /opt/kde/install/darwin/mavericks/clang/shared/general/cmake/share/cmake-3.1/Modules/FeatureSummary.cmake:553 (message): feature_summary() Error: REQUIRED package(s) are missing, aborting CMake run. Call Stack (most recent call first): CMakeLists.txt:106 (feature_summary) CMake Error: The following variables are used in this project, but they are set to NOTFOUND. Please set them or make sure they are set and tested correctly in the CMake files: epoxy_LIBRARY (ADVANCED) linked by target kquickcontrolsaddonsplugin in directory /Users/marko/WC/KDECI-builds/kf5-qt5/kdeclarative/src/qmlcontrols/kquickcontrolsaddons -- Configuring incomplete, errors occurred! ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kdeclarative fails to build on branch master
So, if I do install libepoxy on the OSX/CI system I get this: --- [ 94%] Building CXX object src/qmlcontrols/kquickcontrolsaddons/CMakeFiles/kquickcontrolsaddonsplugin.dir/mimedatabase.cpp.o Building CXX object src/qmlcontrols/kquickcontrolsaddons/CMakeFiles/kquickcontrolsaddonsplugin.dir/plotter.cpp.o /Users/marko/WC/KDECI-builds/kf5-qt5/kdeclarative/src/qmlcontrols/kquickcontrolsaddons/plotter.cpp:23:10: fatal error: 'epoxy/gl.h' file not found #include epoxy/gl.h ^ [ 96%] Scanning dependencies of target kdeclarativetest --- So it looks as if the library cannot be located properly in the system, although it is present: --- $ port contents libepoxy Port libepoxy contains: /opt/local/include/epoxy/gl.h /opt/local/include/epoxy/gl_generated.h /opt/local/include/epoxy/glx.h /opt/local/include/epoxy/glx_generated.h /opt/local/lib/libepoxy.0.dylib /opt/local/lib/libepoxy.dylib /opt/local/lib/pkgconfig/epoxy.pc --- Yet, I’d love to _avoid_ having to install this lib on OSX altogether, as it pulls in all these up to now still unneeded xorg libs: --- $ sudo port activate libepoxy Password: --- Computing dependencies for libepoxy --- Dependencies to be installed: mesa xorg-dri2proto xorg-glproto xorg-libXfixes xorg-fixesproto xorg-xextproto xorg-libXi xorg-inputproto xorg-libXext xorg-libXmu xorg-libXt --- ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kactivities fails to build on branch master
Sorry, this meant of course kactivities and not captivities. :) Looks like a sole boost issue... Gesendet: Freitag, 09. Januar 2015 um 19:03 Uhr Von: Marko Käning mk-li...@email.de An: kde-frameworks-devel kde-frameworks-devel@kde.org Betreff: OSX/CI: captivities fails to build on branch master Linux/CI is quiet, but here on OSX/CI I see this: --- In file included from /opt/local/include/boost/intrusive/circular_slist_algorithms.hpp:24: /opt/local/include/boost/intrusive/detail/common_slist_algorithms.hpp:146:16: error: cannot use 'throw' with exceptions disabled throw; ^ [ 89%] Built target kactivitymanagerd_fileitem_linking_plugin 10 warnings and 1 error generated. [ 90%] make[2]: *** [src/imports/CMakeFiles/kactivitiesextensionplugin.dir/activitiesextensionplugin.cpp.o] Error 1 ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel kactivities.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: captivities fails to build on branch master
Linux/CI is quiet, but here on OSX/CI I see this: --- In file included from /opt/local/include/boost/intrusive/circular_slist_algorithms.hpp:24: /opt/local/include/boost/intrusive/detail/common_slist_algorithms.hpp:146:16: error: cannot use 'throw' with exceptions disabled throw; ^ [ 89%] Built target kactivitymanagerd_fileitem_linking_plugin 10 warnings and 1 error generated. [ 90%] make[2]: *** [src/imports/CMakeFiles/kactivitiesextensionplugin.dir/activitiesextensionplugin.cpp.o] Error 1 kactivities.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: qca-qt5 2.1.0.1
Hi Harald, should perhaps stg like this: --- diff --git a/logical-module-structure b/logical-module-structure index a2ef4e3..14d1f94 100644 --- a/logical-module-structure +++ b/logical-module-structure @@ -294,6 +300,12 @@ kf5-qt5: master, stable-kf5-qt5: master }, +kdesupport/qca: { +stable-qt4: KDE/4.14, +latest-qt4: KDE/4.14, +kf5-qt5: qt5, +stable-kf5-qt5: qt5 +}, playground/pim/akonadi-search: { kf5-qt5: master }, --- get committed to KDE’s build dependencies? Regards, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Some more projects in need of respecting KF5_INSTALL_TARGETS_DEFAULT_ARGS
Hi Alex, Hi Ben, Happy New Year, for a start! :) On 07 Jan 2015, at 19:48 , Alex Merry alex.me...@kde.org wrote: On Tuesday 06 January 2015 23:55:48 Marko Käning wrote: P.P.S.: I realised by now that there are no changes on the production branch, which probably means that there was some change in ECM or wherever provoking these issues... I'm somewhat puzzled, as INSTALL_TARGET_DEFAULT_ARGS should still be set (to the same value as KDE_INSTALL_TARGET_DEFAULT_ARGS) unless you have set KDE_INSTALL_DIRS_NO_DEPRECATED to some TRUE-ish value. If not, that's a bug. Hmmm... Well, off-list I already had contacted Ben because of this issue. See this: Begin forwarded message: On 07 Jan 2015, at 09:12 , David Faure fa...@kde.org wrote: KDEInstallDirs.cmake also has a KDE_INSTALL_TARGETS_DEFAULT_ARGS, which seems more appropriate. I think you're fully correct. Thus I tried KDE_INSTALL_TARGETS_DEFAULT_ARGS for the non-KF5 “systemsettings” project but it lead to the same error as the original INSTALL_TARGETS_DEFAULT_ARGS variable in my initial post. Is this perhaps specific to the CI system? So, I really wonder what’s going on here, as it doesn’t hit all projects! kstars fails, but marble - for instance - still builds fine without the need to mess with these vars! I can't find KDE_INSTALL_DIRS_NO_DEPRECATED in CMakeCache.txt of kstars at all, so I assume that it is not set. Does this mean it IS a bug for kstars and all the other projects after all? Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Some more projects in need of respecting KF5_INSTALL_TARGETS_DEFAULT_ARGS
On 07 Jan 2015, at 20:15 , Christoph Cullmann cullm...@absint.com wrote: will patch that ;=) Thanks for patching this so promptly! :-) ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Linux/CI OSX/CI: kglobalaccel is failing to build on branches stable and master
Hi Martin, On 06 Jan 2015, at 20:00 , Martin Klapetek martin.klape...@gmail.com wrote: Ahh yes, I prepared this and then got carried away by other work, this needs to be done. ok, I will commit these changes then right away. KGlobalAccel got the runtime daemon added into the framework today and is now tier3 (starting from 5.7 release). Ah, ok, I see. (This whole issue rang a bell for me, but I just couldn’t remember where and what I read about it…) Ah humm..that would be my fault, builds fine here. I'll have a look. Great! I'll have a look and fix it. Thanks! Thanks for the investigation! You are welcome. Happy New Year to you! Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kio-extras fails to build on branch master
kio-extras fails due to this: --- CMake Error at network/network/CMakeLists.txt:69 (install): install TARGETS given no LIBRARY DESTINATION for shared library target molletnetwork. --- ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Linux/CI OSX/CI: kglobalaccel is failing to build on branches stable and master
Hi devs, kglobalaccel is currently failing on Jenkins [1]. This can partially be fixed by inserting some dependencies for it: --- diff --git a/dependency-data-kf5-qt5 b/dependency-data-kf5-qt5 index 55b4141..91ac810 100644 --- a/dependency-data-kf5-qt5 +++ b/dependency-data-kf5-qt5 @@ -16,6 +16,13 @@ *: Qt5[stable] *: kdesupport/extra-cmake-modules +# Frameworks, tier1 +frameworks/kglobalaccel: frameworks/kconfig +frameworks/kglobalaccel: frameworks/kcoreaddons +frameworks/kglobalaccel: frameworks/kcrash +frameworks/kglobalaccel: frameworks/kdbusaddons +frameworks/kglobalaccel: frameworks/ki18n + # Frameworks, tier2 frameworks/kauth: frameworks/kcoreaddons frameworks/kcompletion: frameworks/kconfig --- ... yet I wonder whether this is actually wanted!?! I thought that tier 1 frameworks won’t depend on any other framework, but kglobalaccel now seems to be the 1st framework not being in correspondence with this! But well, even with the above lines I run into this after all: --- /Users/marko/WC/KDECI-builds/kglobalaccel/src/runtime/globalshortcutsregistry.cpp:38:10: fatal error: 'kglobalaccel_qws.h' file not found #include kglobalaccel_qws.h ^ 1 error generated. make[2]: *** [src/runtime/CMakeFiles/kglobalaccel5.dir/globalshortcutsregistry.cpp.o] Error 1 --- This seems to be caused by the use of Q_WS_MACX. Replacing it by Q_OS_MAC in src/runtime/globalshortcutsregistry.cpp makes it build the OSX code path: --- diff --git src/runtime/globalshortcutsregistry.cpp src/runtime/globalshortcutsregistry.cpp index 532334a..3f84edb 100644 --- src/runtime/globalshortcutsregistry.cpp +++ src/runtime/globalshortcutsregistry.cpp @@ -30,7 +30,7 @@ #if HAVE_X11 #include kglobalaccel_x11.h -#elif defined(Q_WS_MACX) +#elif defined(Q_OS_MAC) #include kglobalaccel_mac.h #elif defined(Q_WS_WIN) #include kglobalaccel_win.h --- Compiling now gets further, but unfortunately fails: --- In file included from /Users/marko/WC/KDECI-builds/kglobalaccel/src/runtime/globalshortcutsregistry.cpp:34: /Users/marko/WC/KDECI-builds/kglobalaccel/src/runtime/kglobalaccel_mac.h:25:10: fatal error: 'kshortcut.h' file not found #include kshortcut.h ^ --- Ideas? Greets, Marko [1] http://build.kde.org/view/FAILED/job/kglobalaccel_master_qt5/73/console ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: konsole fails to build on branch master
Hi Christoph, On 06 Jan 2015, at 22:10 , Christoph Cullmann cullm...@absint.com wrote: I think the problem is that it should be KF5_INSTALL_TARGETS_DEFAULT_ARGS, the prefix is missing in many cmake files. yep, now I do remember one of your commits from a few days back… So this is the same issue found in many projects then! This would mean, that it DOES make sense to report these issues?! Happy New Year to you, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: konsole fails to build on branch master
Hi Christoph, On 06 Jan 2015, at 22:18 , Christoph Cullmann cullm...@absint.com wrote: Just patch all occurences and avoid sending mails for that, I did only patch the frameworks that I did try to build on the Mac. ok, I personally don’t want to mess in all the projects the CI system is going to fail on in the next few days, which is why I think sending an email listing all the projects currently failing is perhaps the better idea, no? Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
INSTALL_TARGETS_DEFAULT_ARGS (was OSX/CI: .* fails to build on branch .*)
Hi Milian, On 06 Jan 2015, at 21:57 , Milian Wolff m...@milianw.de wrote: On Tuesday 06 January 2015 21:43:03 Marko Käning wrote: CMake Error at src/CMakeLists.txt:178 (install): install TARGETS given no LIBRARY DESTINATION for shared library target konsoleprivate. Again, please stop this spamming. Group the mails, they _all_ show the same error. And it's b/c INSTALL_TARGETS_DEFAULT_ARGS is empty on your machine, for some reason. Go figure out why before sending every compile error you encounter to all lists please. I am sorry, it wasn’t my intention to send spam to any list! As I had run my local OSX/CI system before Christmas w/o any of these messages for more than half a year. This is why I hadn’t expected a change in its functionality and simply reported my findings promptly to the appropriate lists (with one exception pointed out by Albert). I have here tons of frameworks and projects which do NOT show this peculiarity, although they had to be rebuilt as well. OK, I’ll stop reporting for now and check whether there was some major change to build-kde-org’s production branch over Christmas, which I might have missed to merge. Sorry again! A Happy New Year to you, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kde-baseapps fails to build on branch frameworks
CMake Error at keditbookmarks/kbookmarkmodel/CMakeLists.txt:23 (install): install TARGETS given unknown argument BUNDLE. CMake Error at keditbookmarks/CMakeLists.txt:48 (install): install TARGETS given unknown argument BUNDLE”. CMake Error at keditbookmarks/CMakeLists.txt:82 (install): install TARGETS given unknown argument BUNDLE. CMake Error at keditbookmarks/CMakeLists.txt:83 (install): install TARGETS given unknown argument BUNDLE. ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Some more projects in need of respecting KF5_INSTALL_TARGETS_DEFAULT_ARGS
Hi Christoph, I’ve found that these projects do not make use of KF5_INSTALL_TARGETS_DEFAULT_ARGS at the moment: - systemsettings - muon - rocs - libkdegames - kiten - cantor - kolourpaint - libkmahjongg See for details below in order of appearance. I figure this is only the beginning of it and more projects might turn up in the future. Thanks for taking care of these. Regards, Marko P.S.: I’ve tested locally to prepend KF5_” to INSTALL_TARGETS_DEFAULT_ARGS for project systemsettings only, which worked fine and made the project install successfully again here on my OSX/CI system. P.P.S.: I realised by now that there are no changes on the production branch, which probably means that there was some change in ECM or wherever provoking these issues... --- systemsettings: CMake Error at core/CMakeLists.txt:37 (install): install TARGETS given no LIBRARY DESTINATION for shared library target systemsettingsview”. --- muon: CMake Error at libmuon/notifiers/CMakeLists.txt:7 (install): install TARGETS given no LIBRARY DESTINATION for shared library target MuonNotifiers. CMake Error at libmuon/CMakeLists.txt:63 (install): install TARGETS given no LIBRARY DESTINATION for shared library target muonprivate”. --- rocs: CMake Error at libgraphtheory/CMakeLists.txt:106 (install): install TARGETS given no LIBRARY DESTINATION for shared library target rocsgraphtheory”. --- libkdegames: CMake Error at CMakeLists.txt:163 (install): install TARGETS given no LIBRARY DESTINATION for shared library target KF5KDEGames. CMake Error at CMakeLists.txt:222 (install): install TARGETS given no LIBRARY DESTINATION for shared library target KF5KDEGamesPrivate”. --- kiten: CMake Error at lib/CMakeLists.txt:37 (install): install TARGETS given no LIBRARY DESTINATION for shared library target kiten. --- cantor: CMake Error at src/lib/CMakeLists.txt:75 (install): install TARGETS given no LIBRARY DESTINATION for shared library target cantorlibs”. CMake Error at src/CMakeLists.txt:25 (install): install TARGETS given no LIBRARY DESTINATION for shared library target cantor_config. --- kolourpaint: CMake Error at CMakeLists.txt:579 (install): install TARGETS given no LIBRARY DESTINATION for shared library target kolourpaint_lgpl. --- libkmahjongg: CMake Error at CMakeLists.txt:64 (install): install TARGETS given no LIBRARY DESTINATION for shared library target KF5KMahjongglib. CMake Error at CMakeLists.txt:67 (install): install TARGETS given no LIBRARY DESTINATION for shared library target KF5KMahjongglib. --- ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kio placed files erroneously due to missing required backslash in path
Hi, on OSX it looks like some app places files directly in ~/Library, as it misses to insert a backslash in between the standard location ~/Library/Application Support” and the 3 files (user-places.xbel*) to be generated in there: --- $ ls -l1 Library/Application\ Supportuser-places.xbel* -rw--- 1 marko staff 245 Dec 10 03:23 Library/Application Supportuser-places.xbel -rw--- 1 marko staff 512 Dec 10 03:23 Library/Application Supportuser-places.xbel.bak -rw--- 1 marko staff 0 Dec 10 03:23 Library/Application Supportuser-places.xbel.tbcache --- Grep-ing the whole code base to figure out which app caused this shows: --- $ find WC/KDECI-builds/ -name *.cpp -exec grep -H user-places.xbel {} \; WC/KDECI-builds/kbookmarks/autotests/kbookmarktest.cpp:const QString placesFile = datadir + /user-places.xbel; WC/KDECI-builds/kio/src/filewidgets/kfileplacesmodel.cpp:// ~/.local/share/user-places.xbel (according to freedesktop bookmarks spec), and WC/KDECI-builds/kio/src/filewidgets/kfileplacesmodel.cpp:// user-places.xbel will be filled later). (ereslibre) WC/KDECI-builds/kio/src/filewidgets/kfileplacessharedbookmarks.cpp:const QString file = datadir + user-places.xbel; —-- where the last line reveals the culprit as being kio! Fixed in http://commits.kde.org/kio/c5522b6931908d3fd8ad97555a3edf2a3e859b50 Ooops, should I have pushed this through Gerrit before committing? Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kde-baseapps fails to build on branch frameworks
Hi Kevin, Note: That needs Qt 5.4 -- QSignalSpy can take a PMF only since that. I guess that test's code should then be guarded appropriately, so that it builds also on KDE CI's still being used Qt 5.3.2, right? Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: lokalize fails to build on branch frameworks
Hi David, thanks for responding. I'm not sure the lokalize developers read kde-frameworks-devel. I am afraid they are not. I guess they were relying on FindHUNSPELL.cmake from kdelibs 4 which is now in extra-cmake-modules/attic/modules, i.e. it's disabled. You or the lokalize developers should ask kde-buildsystem or look in the archives there (or in the wiki) for the procedure for reviving a cmake module. OK, I will keep this on my TODO list for poking its devs. - e.g. Jeremy?! ;-) Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kde-baseapps fails to build on branch frameworks
/Users/marko/WC/KDECI-builds/kde-baseapps/dolphin/src/tests/kfileitemlistviewtest.cpp:91:16: error: no matching constructor for initialization of 'QSignalSpy' QSignalSpy itemsRemovedSpy(m_model, KFileItemModel::itemsRemoved); ^ ~~ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/include/QtTest/qsignalspy.h:61:14: note: candidate constructor not viable: no known conversion from 'void (KItemModelBase::*)(const KItemRangeList )' to 'const char *' for 2nd argument explicit QSignalSpy(const QObject *obj, const char *aSignal) ^ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/include/QtTest/qsignalspy.h:58:7: note: candidate constructor (the implicit copy constructor) not viable: requires 1 argument, but 2 were provided class QSignalSpy: public QObject, public QListQListQVariant ^ Building CXX object dolphin/src/CMakeFiles/kcm_dolphinviewmodes.dir/settings/viewmodes/viewsettingstab.cpp.o 2 errors generated. [ 50%] make[2]: *** [dolphin/src/tests/CMakeFiles/kfileitemlistviewtest.dir/kfileitemlistviewtest.cpp.o] Error 1 make[1]: *** [dolphin/src/tests/CMakeFiles/kfileitemlistviewtest.dir/all] Error 2 kde-baseapps.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kmime fails to build on branch master
Hi Aleix, On 15 Dec 2014, at 03:01 , Aleix Pol aleix...@kde.org wrote: Thanks for looking into this! :) I won’t look any deeper for now, as it is an issue with the build system itself. I hope this can be clarified once an official OSX build slave is up and running. Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: QSP patch ( was KF5 Update Meeting Minutes 2014-w28 )
Hi David, back to our discussion during this summer: On 10 Jul 2014, at 11:57 , David Faure fa...@kde.org wrote: Wrong approach. XDG is a freedesktop thing which doesn't apply to Mac. This should be fixed the other way around: 1) finding out how Mac OS lets people configure the location for their files, or if this isn't configurable, coming up with QT_* environment variables. 2) patching QStandardPaths to use these 3) setting these vars in the CT scripts Since a fortnight we’re discussing the QStandardPaths patch needed for the OSX/CI system [0] on KDE-MAC [1,2] and it would be good if you’d join this discussion now, as Ben also wants to come up with a patch for Windows very soon. (Ian actually had already included you in his last post [3] to thread [1].) Can we have this discussion on KDE-MAC (or on both lists), as Ian and René aren’t subscribed to K-F-D?! Regards, Marko [0] http://quickgit.kde.org/?p=clones%2Fwebsites%2Fbuild-kde-org%2Fkaning%2Fmp-osx-ci.gita=blobh=27deb1cb8f8f70ea949fc241e049d9b47a10f6dfhb=8126ab8cb976f42a4ebe5d3b7d2cb7a3d42217d1f=patches%2Fqt5%2Fkf5-qt5%2Fpatch-qstandardpaths_mac.cpp.diff [1] http://mail.kde.org/pipermail/kde-mac/2014-December/002590.html [2] http://mail.kde.org/pipermail/kde-mac/2014-December/002705.html [3] http://mail.kde.org/pipermail/kde-mac/2014-December/002737.html ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kio fails to build on branch master
kio is still failing on OSX. Any idea what’g going on with this XML file? On 10 Dec 2014, at 21:02 , Marko Käning mk-li...@email.de wrote: [ 23%] Built target udsentrybenchmark_automoc Scanning dependencies of target ktelnetservice5 Scanning dependencies of target lockingtest Cannot process input: '/Users/marko/WC/KDECI-builds/kio/build/src/ioslaves/http/kcookiejar/org.kde.KCookieServer.xml'. Stop. make[2]: *** [src/ioslaves/http/kcookiejar/kcookieserverinterface.cpp] Error 1 make[1]: *** [src/ioslaves/http/kcookiejar/CMakeFiles/kcookiejar5.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs kio.log.gz___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: libkpeople fails to build on branch master
Hi Aleix, I tried to build all the PIM projects - which you brought recently to CI - on my OSX/CI system. As a consequence I have pushed two additions to dependency-data-kf5-qt5 in [1] and have also spotted a problem with libkpeople. Is its porting to KF5 not yet finished? Greets, Marko [1] http://commits.kde.org/kde-build-metadata/6baa179ab3ee40131d4604175e09924404649ad5 --- -- Detecting CXX compiler ABI info - done CMake Error at /opt/kde/install/darwin/mavericks/clang/shared/general/cmake/share/cmake-3.1/Modules/FindKDE4.cmake:71 (message): ERROR: Could not find KDE4 kde4-config Call Stack (most recent call first): CMakeLists.txt:21 (find_package) CMake Warning (dev) in CMakeLists.txt: No cmake_minimum_required command is present. A line of code such as cmake_minimum_required(VERSION 3.1) should be added at the top of the file. The version specified may be lower if you wish to support older CMake versions for this project. For more information run cmake --help-policy CMP. This warning is for project developers. Use -Wno-dev to suppress it. -- Configuring incomplete, errors occurred! See also /Users/marko/WC/KDECI-builds/libkpeople/build/CMakeFiles/CMakeOutput.log. KDE Continuous Integration Build == Building Project: libkpeople - Branch master libkpeople.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kmime fails to build on branch master
Hi Aleix, it DOES build on Jenkins master [1] with 2 failing tests. So, this is clearly only an issue of how boost is installed and located by the build system on OSX/MacPorts. Greets, Marko [1] http://build.kde.org/view/FAILED/job/kmime_master_qt5/1/ ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: kio fails to build on branch master
Hi Jan, On 13 Dec 2014, at 14:11 , Jan Kundrát j...@kde.org wrote: On Saturday, 13 December 2014 13:55:25 CEST, Marko Käning wrote: Any idea what’g going on with this XML file? One option is to run make with VERBOSE=1 (either asn an env var, or as an argument to make). That way, the output will contain information on what commands are executed, and you should be able to run that single failing command by hand and/or with extra options to see where the problem is. thanks, Jan, for the hint. I’ve found a way how to get this done on the CI system, but when I rebuilt the project I realised that it wasn’t failing anymore… :-/ Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: qt5's QSslSocket can't resolve transfer methods
Turns out that this COULD BE due to the fact that qtdiag grabs the system’s openssl executable in /usr/bin instead of the one installed via MacPorts in /opt/local/bin... --- MVM2:scripts marko$ /usr/bin/openssl OpenSSL version OpenSSL 0.9.8za 5 Jun 2014 --- Since the system’s openssl version is below 1.0 the methods can’t be resolved! Wondering how I could make qtdiag find the one installed in /opt/local/bin, as this --- $ PATH=/opt/local/bin /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/qtdiag $ echo $PATH /Users/marko/bin:/opt/local/bin:/opt/local/sbin:/opt/local/libexec/gnubin:/opt/local/lib/mysql55/bin:/usr/bin:/bin:/usr/sbin:/sbin:/usr/local/bin --- is NOT doing the job, as the PATH env var already has /opt/local/bin first. ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: qt5's QSslSocket can't resolve transfer methods
Hi Ben, On 13 Dec 2014, at 21:39 , Ben Cooksley bcooks...@kde.org wrote: This is where DYLD_LIBRARY_PATH comes in - and it is a variable the CI scripts should already be setting for you. yes, of course! :) Thanks for the reminder! Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: qt5's QSslSocket can't resolve transfer methods
Hi, on my OSX/CI system I realised that qtdiag reports that it cannot resolve certain transfer methods and wonder which system path hasn’t been set correctly here (see below). This e.g. doesn’t allow me to run Trojita on the console [1], but its tests at build time run flawlessly! Any ideas? Greets, Marko [1] http://mail.kde.org/pipermail/kde-mac/2014-December/002634.html --- (This is WITHOUT setting the CI system’s XDG env vars!) $ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/qtdiag QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/var/folders/xd/_025xt7j6dggsjd0_6tczq18gn/T/runtime-marko' QStandardPaths: XDG_RUNTIME_DIR not set, defaulting to '/var/folders/xd/_025xt7j6dggsjd0_6tczq18gn/T/runtime-marko' QSslSocket: cannot resolve TLSv1_1_client_method QSslSocket: cannot resolve TLSv1_2_client_method QSslSocket: cannot resolve TLSv1_1_server_method QSslSocket: cannot resolve TLSv1_2_server_method QSslSocket: cannot resolve SSL_select_next_proto QSslSocket: cannot resolve SSL_CTX_set_next_proto_select_cb QSslSocket: cannot resolve SSL_get0_next_proto_negotiated Qt 5.3.1 (Dec 10 2014, Clang 6.0 (clang-600.0.56) (Apple), 64 bit, debug build) on cocoa little endian/ Mac OS version: 0xb Library info: PrefixPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst DocumentationPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/doc HeadersPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/include LibrariesPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/lib LibraryExecutablesPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/libexec BinariesPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin PluginsPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/plugins ImportsPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/imports Qml2ImportsPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/qml ArchDataPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst DataPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst TranslationsPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/translations ExamplesPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/examples TestsPath: /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/tests Standard paths [*...* denote writable entry]: DesktopLocation: Desktop */Users/marko/Desktop* DocumentsLocation: Documents */Users/marko/Documents* FontsLocation: Fonts ** ApplicationsLocation: Applications ** /Library/Application Support/applications MusicLocation: Music */Users/marko/Music* MoviesLocation: Movies */Users/marko/Videos* PicturesLocation: Pictures */Users/marko/Pictures* TempLocation: TemporaryItems */var/folders/xd/_025xt7j6dggsjd0_6tczq18gn/T* HomeLocation: Home */Users/marko* DataLocation: Application Support */Users/marko/Library/Application Support/Qt Project/qtdiag* /Library/Application Support/Qt Project/qtdiag /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/ CacheLocation: Caches */Users/marko/Library/Caches/Qt Project/qtdiag* /Library/Caches/Qt Project/qtdiag GenericDataLocation: Application Support */Users/marko/Library/Application Support* /Library/Application Support RuntimeLocation: Application Support */var/folders/xd/_025xt7j6dggsjd0_6tczq18gn/T/runtime-marko* ConfigLocation: Preferences */Users/marko/Library/Preferences* /Library/Preferences DownloadLocation: Documents */Users/marko/Downloads* GenericCacheLocation: Caches */Users/marko/Library/Caches* /Library/Caches GenericConfigLocation: Preferences */Users/marko/Library/Preferences* /Library/Preferences ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: breeze fails on branch master
Breeze suddenly fails to build here, as it now depends on X11, although the OSX/CI system globally sets -DCMAKE_DISABLE_FIND_PACKAGE_X11=ON It fails due to xcb: --- /Users/marko/WC/KDECI-builds/breeze/windec/kdecoration2/config/breezedetectwidget.h:43:10: fatal error: 'xcb/xcb.h' file not found #include xcb/xcb.h ^ 1 error generated. make[2]: *** [windec/kdecoration2/CMakeFiles/breezedecoration.dir/config/breezedetectwidget.cpp.o] Error 1 make[2]: *** Waiting for unfinished jobs In file included from /Users/marko/WC/KDECI-builds/breeze/windec/kdecoration2/config/breezeexceptiondialog.cpp:28: /Users/marko/WC/KDECI-builds/breeze/windec/kdecoration2/config/breezedetectwidget.h:43:10: fatal error: 'xcb/xcb.h' file not found #include xcb/xcb.h ^ 1 error generated. make[2]: *** [windec/kdecoration2/CMakeFiles/breezedecoration.dir/config/breezeexceptiondialog.cpp.o] Error 1 ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: breeze fails on branch master
Hi Martin, On 12 Dec 2014, at 08:02 , Martin Gräßlin mgraess...@kde.org wrote: X11 and xcb are not the same. oh, I see. :) On 12 Dec 2014, at 08:12 , Martin Gräßlin mgraess...@kde.org wrote: I hope this fixes it: http://commits.kde.org/breeze/d35fb88cd9d9a678a170f01d9643c8956c38aee5 Yep, that did it. Thanks for taking care of this. Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kio fails to build on branch master
[ 23%] Built target udsentrybenchmark_automoc Scanning dependencies of target ktelnetservice5 Scanning dependencies of target lockingtest Cannot process input: '/Users/marko/WC/KDECI-builds/kio/build/src/ioslaves/http/kcookiejar/org.kde.KCookieServer.xml'. Stop. make[2]: *** [src/ioslaves/http/kcookiejar/kcookieserverinterface.cpp] Error 1 make[1]: *** [src/ioslaves/http/kcookiejar/CMakeFiles/kcookiejar5.dir/all] Error 2 make[1]: *** Waiting for unfinished jobs kio.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: Anyone interested in test results of all KF5 frameworks and many KF5 apps?
Hi, I am soon about to finalise the first complete rebuild of KF5 on my OSX/CI system NOW INCLUDING ALL EXISTING TESTS ! Anyone interested in all those test results for FWs apps with test failures? Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: GenericConfigLocation missing in qtpaths tool
Hi David Ben, On 06 Dec 2014, at 21:13 , David Faure fa...@kde.org wrote: Fixed 5 months ago: commit 3b102b026b6fdd8dfbc284741d8df7c1d6dcc9d3 Author: David Faure david.fa...@kdab.com Date: Sun Jul 6 15:30:58 2014 +0200 qtpaths: add missing GenericConfigLocation, exists since Qt 5.2.0 Change-Id: I4d4ef2e0f33896984f436df6c9a033dd8985 Reviewed-by: Sune Vuorela s...@vuorela.dk Reviewed-by: Thiago Macieira thiago.macie...@intel.com $ git tag --contains 3b102b026b6fdd8dfbc284741d8df7c1d6dcc9d3 v5.3.2 v5.4.0-alpha1 v5.4.0-beta1 v5.4.0-rc1 oh, I thought that I am running a rather recent Qt5 installation on the OSX/CI system. I built at the end of November, but I see now that it is 5.3.1: --- $ gtdiag Qt 5.3.1 (Nov 21 2014, Clang 6.0 (clang-600.0.54) (Apple), 64 bit, debug build) on cocoa little endian/ $ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/qtpaths --version qtpaths 1.0 --- That explains it. Greets, Marko signature.asc Description: Message signed with OpenPGP using GPGMail ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: phonon fails to INSTALL on branch master
On 27 Nov 2014, at 08:59 , Marko Käning mk-li...@email.de wrote: Phonon built fine, but failed to install for some reason, as a file must be missing: --- -- Installing: /Users/marko/WC/KDECI-builds/phonon/local-inst/opt/kde/install/darwin/mavericks/clang/kf5-qt5/kdesupport/phonon/phonon/inst/include/phonon4qt5/phonon/VolumeSlider CMake Error at includes/cmake_install.cmake:88 (file): file INSTALL cannot find /Users/marko/WC/KDECI-builds/phonon/includes/old/Phonon/AbstractAudioOutput. Call Stack (most recent call first): cmake_install.cmake:69 (include) make: *** [install] Error 1 KDE Continuous Integration Build Just for the record, I saw this error once again!!! Building phonon a 2nd time it was successful though. ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: One kde-baseapps test fails
This is one of my first test results on OSX: FAIL! : KonqPopupMenuTest::testSubDirectory() Received a fatal error. Full test results are attached. kde-baseapps-test-results.tar.bz2 Description: BZip2 compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: All Kate tests failing
All of Kate’s tests are failing on OSX/CI with this message: Unable to find executable: session_test Since other projects are able to run tests successfully, I gather that this is a kate-specific issue. Test results attached. kate-test-results.tar.bz2 Description: BZip2 compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: KFileMetaData fails to build on branch master
Building CXX object src/CMakeFiles/KF5FileMetaData.dir/simpleextractionresult.cpp.o Generating moc_office2007extractortest.cpp Generating moc_office2007extractor.cpp [ 41%] [ 41%] Built target officeextractortest_automoc Building CXX object src/CMakeFiles/KF5FileMetaData.dir/typeinfo.cpp.o [ 43%] Building CXX object src/CMakeFiles/KF5FileMetaData.dir/usermetadata.cpp.o /Users/marko/WC/KDECI-builds/kfilemetadata/src/usermetadata.cpp:140:13: error: use of undeclared identifier 'errno' return (errno != ENODATA); ^ /Users/marko/WC/KDECI-builds/kfilemetadata/src/usermetadata.cpp:140:22: error: use of undeclared identifier 'ENODATA' return (errno != ENODATA); ^ /Users/marko/WC/KDECI-builds/kfilemetadata/src/usermetadata.cpp:156:13: error: use of undeclared identifier 'errno' return (errno == ENOTSUP); ^ /Users/marko/WC/KDECI-builds/kfilemetadata/src/usermetadata.cpp:156:22: error: use of undeclared identifier 'ENOTSUP' return (errno == ENOTSUP); ^ 4 errors generated. make[2]: *** [src/CMakeFiles/KF5FileMetaData.dir/usermetadata.cpp.o] Error 1 make[2]: *** Waiting for unfinished jobs make[1]: *** [src/CMakeFiles/KF5FileMetaData.dir/all] Error 2 make: *** [all] Error 2 KDE Continuous Integration Build == Building Project: kfilemetadata - Branch master ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: KFileMetaData fails to build on branch master
Hi David, On 06 Dec 2014, at 14:53 , David Faure fa...@kde.org wrote: On Saturday 06 December 2014 13:08:19 Marko Käning wrote: /Users/marko/WC/KDECI-builds/kfilemetadata/src/usermetadata.cpp:140:13: error: use of undeclared identifier 'errno' return (errno != ENODATA); ^ Fixed. thanks for taking care of this one. BUT, now I am back to the old taglib issue reported in September to KFileMetaData’s developers… :-( Will create a new post. Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
OSX/CI: kfilemetadata fails to build on branch master
The inclusion of MacPorts’ taglib fails on OSX. See the below discussions about this problem from September. kfilemetadata.log.gz Description: GNU Zip compressed data Begin forwarded message: From: Marko Käning mk-li...@email.de Subject: taglib inclusion on OSX Date: 13 Sep 2014 22:54:07 GMT+2 To: j...@jriddell.org, handa.v...@gmail.com Cc: bhus...@gmail.com Hi Jonathan Vishesh, Ben suggested to send our below conversation to you guys, as I am seeing problems with including taglib on OSX for plasma-mediacenter, but also for the project KFileMetaData maintained by you two. Greets, Marko P.S.: Separately I have already contacted PMC’s Bhushan Shah, who asked me to test building PMC while at Akademy, which I could do after some patching. We need a solution which does not require patching taglib’s sources. From: Ben Cooksley bcooks...@kde.org Subject: Re: Plasma Media Center - Build Success on OSX Date: 13 Sep 2014 22:09:33 GMT+2 To: Marko Käning mk-li...@email.de On Sun, Sep 14, 2014 at 6:06 AM, Marko Käning mk-li...@email.de wrote: Hi Ben, Hi Marko, I am forwarding what I wrote to Bhushan earlier today. I wonder whether you might have a hint as to what is going on here. On MacPorts exists the an non-patched taglib version 1.9.1. Although I patched it regarding the pc files locally I didn’t succeed building plasma-mediacenter without patching is as well. This problem occurs just as well for kfilemetadata, where I also had to add /opt/local/include to the include directories. It seems as if TAGLIB_INCLUDES is not generated from any of taglib's pc files, right? TAGLIB_INCLUDES will probably be set by a FindTaglib.cmake or TaglibConfig.cmake file - either of which could potentially use the pkgconfig *.pc files provided by Taglib. This issue is a fairly routine case of include statements being inconsistent with where things are being installed. It only works for him and others because Taglib is usually installed in /usr on Linux (and if it isn't there, it probably co-exists with kdelibs/frameworks which add the $prefix/include directory to the include_directories). The proper way to fix this is to either: 1) Adjust the PMC code to include the taglib headers without the taglib/ prefix. 2) Adjust the CMake code which locates taglib to add the $prefix/include directory in addition to the $prefix/include/taglib directory. Greets, Marko Thanks, Ben From: Marko Käning mk-li...@email.de Subject: Fwd: Plasma Media Center - Build Success on OSX Date: 13 Sep 2014 20:06:34 GMT+2 To: Ben Cooksley bcooks...@kde.org Hi Ben, I am forwarding what I wrote to Bhushan earlier today. I wonder whether you might have a hint as to what is going on here. On MacPorts exists the an non-patched taglib version 1.9.1. Although I patched it regarding the pc files locally I didn’t succeed building plasma-mediacenter without patching is as well. This problem occurs just as well for kfilemetadata, where I also had to add /opt/local/include to the include directories. It seems as if TAGLIB_INCLUDES is not generated from any of taglib's pc files, right? Greets, Marko Begin forwarded message: From: Marko Käning mk-li...@email.de Subject: Plasma Media Center - Build Success on OSX Date: 13 Sep 2014 15:21:57 GMT+2 To: bhus...@gmail.com Hi Bhushan, just for your information, once I appended my local prefix here: --- message(STATUS TAGLIB_INCLUDES = ${TAGLIB_INCLUDES}) include_directories( ${TAGLIB_INCLUDES} /opt/local/include ) --- I was able to ***successfully*** build PMC on OSX. :-) For that to work I had to disable the deps to baloo and mockcpp, of course. :) Well, checking the contents of TAGLIB_INCLUDES - as shown above - I got this: --- TAGLIB_INCLUDES = ;/opt/local/include/taglib --- But your code prepends another taglib subfolder to every file included. This is indeed very weird and I don’t know why this is happening, as the pc files of my current taglib installation do NOT have that path appended, but taglib-config tells me otherwise --- $ taglib-config --cflags -I/opt/local/include/taglib $ grep Cflags /opt/local/lib/pkgconfig/taglib.pc Cflags: -I/opt/local/include --- What are you seeing on your Linux system here? (Actually I have manipulated those Cflags settings myself locally, as version 1.9.1 actually has “taglib” appended. See below! Looks like those changes in the pc files aren’t enough to make this work. I haven’t dived into the taglib-config code now, I must say.) I am puzzled and wonder what’s going on here… Has taglib messed up or is there some CMake problem in KDE’s build system here? Greets, Marko P.S.: BTW, I've run into the same problem with kfilemetadata - now that I had to install taglib on my OSX/CI system. :-/ That’s why I am interested in finding a solution for this little, but disturbing
GenericConfigLocation missing in qtpaths tool
Hi folks, there is a ConfigLocation path type: --- $ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/qtpaths --paths ConfigLocation /Users/marko/Library/Preferences $ /opt/kde/install/darwin/mavericks/clang/kf5-qt5/qt5/inst/bin/qtpaths --paths GenericConfigLocation Unknown location: GenericConfigLocation --- but for the path type GenericConfigLocation the result is _unknown_. As I stumbled over this again I wanted to ask whether it is just a peculiarity of the qtpaths tool? Greets, Marko ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: OSX/CI: phonon fails to INSTALL on branch master
Hi Ben, On 27 Nov 2014, at 10:40 , Harald Sitter sit...@kde.org wrote: On Thu, Nov 27, 2014 at 8:59 AM, Marko Käning mk-li...@email.de wrote: Phonon built fine, but failed to install for some reason, as a file must be missing: I think that needs some more investigation. I can not reproduce this nor can I even think of how that would happen :/ as Harald could not reproduce this on Linux I wondered what I could do to diagnose this issue further. Well - since I had no other ideas - I just tried installing once again and see: IT INSTALLED FINE AFTER ALL !!! (The current log is attached.) What could have been going on there? Perhaps some destination issue? Any hints for me? Greets, Marko phonon.log.gz Description: GNU Zip compressed data ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: KDE-CI IRC meeting - Your possible KDE contributions in non-C++
Hi Mario, thanks for organizing this! Prospective agenda for the IRC meeting: - Ben: Short introduction to KDE CI - Everybody: Short introduction incl. their skills and wishes for KDE CI - Ben: What needs to be changed - Everybody: Work on a roadmap and distribute work till next meeting - Everybody: Determine date for the next IRC meeting I could give a short presentation about my OSX/CI system and describe where I'd need support. Hey, we should go through our list of topics for the IRC meetings, which we created months ago! :-) Greets, Marko P.S.: I already see that Jeremy has a shifted time-slot wrt me. I'll have to check whether I can make it 3pm my time some day. The bigger problem will be to synchronize with Ben... ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120969: Fix build on OSX due to missing XDR functions.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120969/#review70715 --- Ship it! Ship It! - Marko Käning On Nov. 4, 2014, 4:58 p.m., Mathias Tillman wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120969/ --- (Updated Nov. 4, 2014, 4:58 p.m.) Review request for KDE Frameworks and David Faure. Repository: kio-extras Description --- Due to some missing XDR functions on OSX (used to decode unsigned long long), it failed to compile. This fixes that by doing some checks for the existence of the 64-bit datatype functions (for some reason XDR have 5 different functions that do the same thing). This patch also merges the mount3 rpc code with the nfs3 code as well as remove the unused rpc mount2 code. Diffs - nfs/rpc_mnt3.h cba4ff656a3d524f39cb7f607b8b4b331b41a0c5 nfs/rpc_mnt2_xdr.c b380f4f93853ba54112a68ddf851f8f04b251312 nfs/rpc_mnt2.x 4aaf97de9ade2655cd48ae460ba24ee322b57183 nfs/nfsv3.cpp 368a61febc778688507ffb3954caaf69fc18d371 nfs/rpc_nfs2_prot.x cd21123c7b57040d97245f22038153945abe88ee nfs/rpc_nfs2_prot.h 62ab3056a51cc0e89536081c9d5741ff136dc804 nfs/rpc_nfs3_prot_xdr.c 57f4a8546ae697f41f98efb24c01b6de739d380f nfs/rpc_nfs3_prot.x fba417c8cd6edb955e018a89dd38942f1e788c69 nfs/rpc_nfs3_prot.h 76b2f5becef9cbdc9fd3db08c540c065a43a89a6 nfs/rpc_nfs2_prot_xdr.c fb63b8ec2329730ef55c8f068e8abf47e865173b nfs/rpc_mnt3_xdr.c 2016e688770e3753affe20923a8515f2427c5447 nfs/rpc_mnt2.h ed21256c4754ce861ff1f5ea5ddc4ca224a7da7e nfs/kio_nfs.h 9943daf670e665121ea51419ea77599ec4d8d4f5 nfs/CMakeLists.txt 2ed2186e113dad9e9c178cbac680d7c96dff789a Diff: https://git.reviewboard.kde.org/r/120969/diff/ Testing --- I haven't personally tried to compile it on OSX, but Marko Käning (the one who reported it, https://www.mail-archive.com/kde-frameworks-devel@kde.org/msg19770.html) confirmed that it's now compiling fine on OSX. Thanks, Mathias Tillman ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel