Re: Review Request 122493: Use math.h round rather than C++11 std::lround.
On Feb. 9, 2015, 12:59 a.m., Mark Gaiser wrote: -1 You use C-Style casts. Oke, the frameworks coding style doesn't seem to explicitly forbid it (casts are not mentioned), but if i recall correctly we use the Qt style + some of our own which means we should obey the Qt style [1] which does mention casts and forbids the C-Sytle cast. Secondly, you seem to be trying to get this working on a compiler that isn't supported [2]. If one of those compilers have issues with std::lround (which i seriously doubt) then we should perhaps look into replacing it with qRound. However, Qt is also moving away from it's own algorithms in favor of the stl ones so we should stick to the std:: versions. Cheers [1] http://qt-project.org/wiki/Qt_Coding_Style [2] https://community.kde.org/Frameworks/Policies#Frameworks_compiler_requirements_and_C.2B.2B11 Ivan Čukić wrote: I'm torn on this one. This is a minor edit (mainly thinking about the potential qRound version) that would fix the current master to work with an old platform. Without having any downsides (apart from possible future depreciacion of qRound). But, on the other hand, something else will probably be brought in soon that will require some more essential c++11 things, and break on osx 10.7. So, one wonders whether a patch that just prolongs the inevitable has a purpose. :) Kevin Funk wrote: Not sure if it's worth discussion this issue a lot... `qRound` isn't deprecated yet, used a lot throughout frameworks, and it doesn't look like it will be deprecated (I've yet to see a discussion on the Qt ML about this). Yet `qRound` is way more pleasant to read than `static_castint(...)`. ok for using qRound for now if it fixes build on old platforms - Marco --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/#review75648 --- On Feb. 8, 2015, 11:12 p.m., Jeremy Whiting wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/ --- (Updated Feb. 8, 2015, 11:12 p.m.) Review request for KDE Frameworks and Marco Martin. Repository: kdeclarative Description --- Use math.h round 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: strange problem with KConfig
On Monday 09 February 2015, Matthew Dawson wrote: However, I think the bug can be worked around in the short term for the kcm by having the KSharedConfigs in question use only the file they are pointed at. I think that is actually the correct behaviour, as those files are not supposed to use the global or system configuration files. I've attached a (relatively untested) sample patch. It fixes the bug, but I'm not sure if it allows the kcm to continue functioning as expected. to me the patch seems fine (Hrvoje tested it for his use case, seems also to work correctly). so for now I would go with it, can you push? -- Marco Martin ___ 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/ --- (Updated Feb. 9, 2015, 8:31 a.m.) Review request for KDE Frameworks and Marco Martin. Changes --- Changed to qRound Summary (updated) - Use qRound rather than C++11 std::lround. Repository: kdeclarative Description (updated) --- Use qRound rather than C++11 std::lround. Removes dependency on C++11. Diffs (updated) - 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 122495: CMake nitpicking on KDiagram
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122495/#review75704 --- Ship it! Ship It! - Friedrich W. H. Kossebau On Feb. 9, 2015, 3:09 vorm., Aleix Pol Gonzalez wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122495/ --- (Updated Feb. 9, 2015, 3:09 vorm.) Review request for KDE Frameworks and Friedrich W. H. Kossebau. Repository: kdiagram Description --- Mark Qt5::Widgets public for both KCharts and KGantt. Remove SHARED from add_library, let cmake use the default (which is SHARED, but the user can configure). Fix some indentation. Remove redundant dependencies: if we depend on Qt5::Widgets, we're already pulling Qt5::Gui. Diffs - src/KChart/CMakeLists.txt 06f3846 src/KGantt/CMakeLists.txt 25d198f CMakeLists.txt 76a7c50 Diff: https://git.reviewboard.kde.org/r/122495/diff/ Testing --- Still builds. Thanks, Aleix Pol Gonzalez ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 122481: Fix use of environ for OS X
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122481/ --- (Updated Feb. 9, 2015, 3:17 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks, David Faure, Marko Käning, and René J.V. Bertin. Repository: kio Description --- On OS X 10.7 environ needs to be #defined as _NSGetEnviron() from crt_externs.h according to man environ. This is also required for kio to build on OS X 10.7 Diffs - src/widgets/kurlcompletion.cpp 1871c49a936e2ca322286e23ad6fe976ae2c7044 Diff: https://git.reviewboard.kde.org/r/122481/diff/ Testing --- kio builds with this patch on OS X 10.7 and 10.10 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 122487: Reorder member variables to reduce memory usage.
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122487/ --- (Updated Feb. 9, 2015, 5:04 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks. Repository: kcoreaddons Description --- The KJobPrivate one saves 24 bytes per instance, the others 8 bytes (on a 64 bit system). Diffs - src/lib/io/kdirwatch_p.h f8cc2bcd5b9a8c65b409873ac78455fbb7db2405 src/lib/jobs/kjob.cpp de3b8d7c8985dcc21ef9f52fba47361e9675f192 src/lib/jobs/kjob_p.h 8bacefecc4dc8d2679a48b224837e0c01516d601 src/lib/plugin/kpluginloader.cpp 64d7f0cec37ef51a76a3a60637822810cab398e6 Diff: https://git.reviewboard.kde.org/r/122487/diff/ Testing --- ``` Struct Entry (DW_TAG_class_type, offset 393887) is sub-optimally packed: 74/88 bytes, 592/704 bits, optimal size is: 80 bytes class KDirWatchPrivate::Entry // location: /k/kde5/src/kcoreaddons/src/lib/io/kdirwatch_p.h:83 { time_t m_ctime; // member offset: 0, size: 8, alignment: 8 int m_nlink; // member offset: 8, size: 4, alignment: 4 // 4 byte(s) padding ino_t m_ino; // member offset: 16, size: 8, alignment: 8 entryStatus m_status; // member offset: 24, size: 4, alignment: 4 entryMode m_mode; // member offset: 28, size: 4, alignment: 4 bool isDir; // member offset: 32, size: 1, alignment: 1 // 7 byte(s) padding QListKDirWatchPrivate::Client* m_clients; // member offset: 40, size: 8, alignment: 8 QListKDirWatchPrivate::Entry* m_entries; // member offset: 48, size: 8, alignment: 8 QString path; // member offset: 56, size: 8, alignment: 8 int msecLeft; // member offset: 64, size: 4, alignment: 4 int freq; // member offset: 68, size: 4, alignment: 4 bool dirty; // member offset: 72, size: 1, alignment: 1 // 3 byte(s) padding int wd; // member offset: 76, size: 4, alignment: 4 QListQString m_pendingFileChanges; // member offset: 80, size: 8, alignment: 8 }; // size: 88, alignment: 8 Struct KJobPrivate (DW_TAG_class_type, offset 1112242) is sub-optimally packed: 95/120 bytes, 760/960 bits, optimal size is: 96 bytes class KJobPrivate // location: /k/kde5/src/kcoreaddons/src/lib/jobs/kjob_p.h:36 { __vtbl_ptr_type* _vptr.KJobPrivate; // member offset: 0, size: 8, alignment: 8 KJob* q_ptr; // member offset: 8, size: 8, alignment: 8 KJobUiDelegate* uiDelegate; // member offset: 16, size: 8, alignment: 8 int error; // member offset: 24, size: 4, alignment: 4 // 4 byte(s) padding QString errorText; // member offset: 32, size: 8, alignment: 8 Unit progressUnit; // member offset: 40, size: 4, alignment: 4 // 4 byte(s) padding QMapKJob::Unit, long long unsigned int processedAmount; // member offset: 48, size: 8, alignment: 8 QMapKJob::Unit, long long unsigned int totalAmount; // member offset: 56, size: 8, alignment: 8 long unsigned int percentage; // member offset: 64, size: 8, alignment: 8 bool suspended; // member offset: 72, size: 1, alignment: 1 // 3 byte(s) padding Capabilities capabilities; // member offset: 76, size: 4, alignment: 4 QTimer* speedTimer; // member offset: 80, size: 8, alignment: 8 bool isAutoDelete; // member offset: 88, size: 1, alignment: 1 // 7 byte(s) padding QEventLoop* eventLoop; // member offset: 96, size: 8, alignment: 8 QEventLoopLocker eventLoopLocker; // member offset: 104, size: 8, alignment: 8 bool isFinished; // member offset: 112, size: 1, alignment: 1 // 7 byte(s) padding }; // size: 120, alignment: 8 Struct KPluginLoaderPrivate (DW_TAG_class_type, offset 1584914) is sub-optimally packed: 37/48 bytes, 296/384 bits, optimal size is: 40 bytes class KPluginLoaderPrivate // location: /k/kde5/src/kcoreaddons/src/lib/plugin/kpluginloader.cpp:37 { KPluginLoader* q_ptr; // member offset: 0, size: 8, alignment: 8 QString const name; // member offset: 8, size: 8, alignment: 8 quint32 pluginVersion; // member offset: 16, size: 4, alignment: 4 // 4 byte(s) padding QString errorString; // member offset: 24, size: 8, alignment: 8 bool pluginVersionResolved; // member offset: 32, size: 1, alignment: 1 // 7 byte(s) padding QPluginLoader* loader; // member offset: 40, size: 8, alignment: 8 }; // size: 48, alignment: 8 ``` Thanks, Volker Krause ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 122495: CMake nitpicking on KDiagram
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122495/ --- (Updated Feb. 9, 2015, 4:11 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks and Friedrich W. H. Kossebau. Repository: kdiagram Description --- Mark Qt5::Widgets public for both KCharts and KGantt. Remove SHARED from add_library, let cmake use the default (which is SHARED, but the user can configure). Fix some indentation. Remove redundant dependencies: if we depend on Qt5::Widgets, we're already pulling Qt5::Gui. Diffs - src/KChart/CMakeLists.txt 06f3846 src/KGantt/CMakeLists.txt 25d198f CMakeLists.txt 76a7c50 Diff: https://git.reviewboard.kde.org/r/122495/diff/ Testing --- Still builds. Thanks, Aleix Pol Gonzalez ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120185: Look for kdesu in the correct location
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120185/ --- (Updated Feb. 9, 2015, 5:41 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks and David Faure. Bugs: 338755 https://bugs.kde.org/show_bug.cgi?id=338755 Repository: kio Description --- kdesu is installed in libexec/ look for it there first. I left the findExecutable search as a backup. Is looking in CMAKE_INSTALL_FULL_LIBEXECDIR correct? Or will kde-cli-tools be installed in libexec/kf5? Insert 'kdesu' at the end to show a nicer error. If we leave this part out the error is Could not launch 'root' which is somewhat correct but not as easy to figure out as Could not launch 'kdesu' Also added an unrelated QFile::decodeName() call. Diffs - autotests/krununittest.cpp b1da9aa src/core/desktopexecparser.cpp 9510697 Diff: https://git.reviewboard.kde.org/r/120185/diff/ Testing --- Created .desktop file with X-KDE-SubstituteUID=true Now I can launch it as root and when I remove kdesu I got a normal error message. Unit test no longer skips because kdesu is not found. Thanks, Maarten De Meyer ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/#review75726 --- +1 - Mark Gaiser On feb 9, 2015, 5:51 p.m., Àlex Fiestas wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated feb 9, 2015, 5:51 p.m.) Review request for KDE Frameworks. Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing --- Besides tests, tried smb kioslave and it worked great. Thanks, Àlex Fiestas ___ 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/ --- (Updated Feb. 9, 2015, 7:26 p.m.) Status -- This change has been marked as submitted. 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: Fixing and regulating certain types of search fields across KF5 apps
Eike, On Mon, Feb 9, 2015 at 1:03 PM, Eike Hein h...@kde.org wrote: Hi, we currently have numerous dynamic search fields across KDE apps that, for performance reasons, disregard input shorter than a particular QString::length(). Examples are KMail and some things in KRunner. It's a common pattern. This is very English-centric behavior; many scripts are more complex than this, in particular CJK ones. In Korean, a single character may be composed of multiple letters and form a complex syllable. Most Korean given names are two characters, many useful words are only one character. The situation in Japanese is similar - it's written using a mix of syllabaries and ideograms. The check for 3 has good intentions but breaks real use cases for these users, such as finding mail by name. Addressing this algorithmically is not all that hard. When normalizing to Unicode NFD, those complex Korean characters turn into a sequence of combining characters for the indi- vidual letters. Syllabic and ideographic code points are assigned the character class Lo in Unicode, which is info available via QChar. An NFD pass + counting Lo chars as 2 would be simple and effective enough for this purpose. The way I'd like to see this fixed is placing an API like isMinimumSearchableLength(QString) somewhere in KF5 and to start using it everywhere. Realistically developers won't find this on their own, but it means I could go around and patch up all these places without copying code around. That's a great idea imo. Which framework would be a good place for this? Without giving it much thought I would think either KI18n or Sonnet. Cheers, Eike ___ 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
Re: Review Request 122493: Use math.h round rather than C++11 std::lround.
On Feb. 9, 2015, 12:59 a.m., Mark Gaiser wrote: -1 You use C-Style casts. Oke, the frameworks coding style doesn't seem to explicitly forbid it (casts are not mentioned), but if i recall correctly we use the Qt style + some of our own which means we should obey the Qt style [1] which does mention casts and forbids the C-Sytle cast. Secondly, you seem to be trying to get this working on a compiler that isn't supported [2]. If one of those compilers have issues with std::lround (which i seriously doubt) then we should perhaps look into replacing it with qRound. However, Qt is also moving away from it's own algorithms in favor of the stl ones so we should stick to the std:: versions. Cheers [1] http://qt-project.org/wiki/Qt_Coding_Style [2] https://community.kde.org/Frameworks/Policies#Frameworks_compiler_requirements_and_C.2B.2B11 I'm torn on this one. This is a minor edit (mainly thinking about the potential qRound version) that would fix the current master to work with an old platform. Without having any downsides (apart from possible future depreciacion of qRound). But, on the other hand, something else will probably be brought in soon that will require some more essential c++11 things, and break on osx 10.7. So, one wonders whether a patch that just prolongs the inevitable has a purpose. :) - Ivan --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/#review75648 --- On Feb. 8, 2015, 11:12 p.m., Jeremy Whiting wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/ --- (Updated Feb. 8, 2015, 11:12 p.m.) Review request for KDE Frameworks and Marco Martin. Repository: kdeclarative Description --- Use math.h round 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 math.h round rather than C++11 std::lround.
On Feb. 9, 2015, 12:59 a.m., Mark Gaiser wrote: -1 You use C-Style casts. Oke, the frameworks coding style doesn't seem to explicitly forbid it (casts are not mentioned), but if i recall correctly we use the Qt style + some of our own which means we should obey the Qt style [1] which does mention casts and forbids the C-Sytle cast. Secondly, you seem to be trying to get this working on a compiler that isn't supported [2]. If one of those compilers have issues with std::lround (which i seriously doubt) then we should perhaps look into replacing it with qRound. However, Qt is also moving away from it's own algorithms in favor of the stl ones so we should stick to the std:: versions. Cheers [1] http://qt-project.org/wiki/Qt_Coding_Style [2] https://community.kde.org/Frameworks/Policies#Frameworks_compiler_requirements_and_C.2B.2B11 Ivan Čukić wrote: I'm torn on this one. This is a minor edit (mainly thinking about the potential qRound version) that would fix the current master to work with an old platform. Without having any downsides (apart from possible future depreciacion of qRound). But, on the other hand, something else will probably be brought in soon that will require some more essential c++11 things, and break on osx 10.7. So, one wonders whether a patch that just prolongs the inevitable has a purpose. :) Not sure if it's worth discussion this issue a lot... `qRound` isn't deprecated yet, used a lot throughout frameworks, and it doesn't look like it will be deprecated (I've yet to see a discussion on the Qt ML about this). Yet `qRound` is way more pleasant to read than `static_castint(...)`. - Kevin --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/#review75648 --- On Feb. 8, 2015, 11:12 p.m., Jeremy Whiting wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122493/ --- (Updated Feb. 8, 2015, 11:12 p.m.) Review request for KDE Frameworks and Marco Martin. Repository: kdeclarative Description --- Use math.h round 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 122487: Reorder member variables to reduce memory usage.
On Feb. 8, 2015, 8:45 p.m., David Faure wrote: Which magic tool did you come up with this time, to detect such things? :-) Aleix Pol Gonzalez wrote: +1 :D I too want to know! It's still very much WIP, the code is in kde:scratch/vkrause/elf-dissector but be aware that it still generates arbitrarily wrong output, there's still many cases it's not detecting/handling correctly. And it'll likely crash if you are not on x86_64 with a full debug build. - Volker --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122487/#review75629 --- On Feb. 9, 2015, 5:04 p.m., Volker Krause wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122487/ --- (Updated Feb. 9, 2015, 5:04 p.m.) Review request for KDE Frameworks. Repository: kcoreaddons Description --- The KJobPrivate one saves 24 bytes per instance, the others 8 bytes (on a 64 bit system). Diffs - src/lib/io/kdirwatch_p.h f8cc2bcd5b9a8c65b409873ac78455fbb7db2405 src/lib/jobs/kjob.cpp de3b8d7c8985dcc21ef9f52fba47361e9675f192 src/lib/jobs/kjob_p.h 8bacefecc4dc8d2679a48b224837e0c01516d601 src/lib/plugin/kpluginloader.cpp 64d7f0cec37ef51a76a3a60637822810cab398e6 Diff: https://git.reviewboard.kde.org/r/122487/diff/ Testing --- ``` Struct Entry (DW_TAG_class_type, offset 393887) is sub-optimally packed: 74/88 bytes, 592/704 bits, optimal size is: 80 bytes class KDirWatchPrivate::Entry // location: /k/kde5/src/kcoreaddons/src/lib/io/kdirwatch_p.h:83 { time_t m_ctime; // member offset: 0, size: 8, alignment: 8 int m_nlink; // member offset: 8, size: 4, alignment: 4 // 4 byte(s) padding ino_t m_ino; // member offset: 16, size: 8, alignment: 8 entryStatus m_status; // member offset: 24, size: 4, alignment: 4 entryMode m_mode; // member offset: 28, size: 4, alignment: 4 bool isDir; // member offset: 32, size: 1, alignment: 1 // 7 byte(s) padding QListKDirWatchPrivate::Client* m_clients; // member offset: 40, size: 8, alignment: 8 QListKDirWatchPrivate::Entry* m_entries; // member offset: 48, size: 8, alignment: 8 QString path; // member offset: 56, size: 8, alignment: 8 int msecLeft; // member offset: 64, size: 4, alignment: 4 int freq; // member offset: 68, size: 4, alignment: 4 bool dirty; // member offset: 72, size: 1, alignment: 1 // 3 byte(s) padding int wd; // member offset: 76, size: 4, alignment: 4 QListQString m_pendingFileChanges; // member offset: 80, size: 8, alignment: 8 }; // size: 88, alignment: 8 Struct KJobPrivate (DW_TAG_class_type, offset 1112242) is sub-optimally packed: 95/120 bytes, 760/960 bits, optimal size is: 96 bytes class KJobPrivate // location: /k/kde5/src/kcoreaddons/src/lib/jobs/kjob_p.h:36 { __vtbl_ptr_type* _vptr.KJobPrivate; // member offset: 0, size: 8, alignment: 8 KJob* q_ptr; // member offset: 8, size: 8, alignment: 8 KJobUiDelegate* uiDelegate; // member offset: 16, size: 8, alignment: 8 int error; // member offset: 24, size: 4, alignment: 4 // 4 byte(s) padding QString errorText; // member offset: 32, size: 8, alignment: 8 Unit progressUnit; // member offset: 40, size: 4, alignment: 4 // 4 byte(s) padding QMapKJob::Unit, long long unsigned int processedAmount; // member offset: 48, size: 8, alignment: 8 QMapKJob::Unit, long long unsigned int totalAmount; // member offset: 56, size: 8, alignment: 8 long unsigned int percentage; // member offset: 64, size: 8, alignment: 8 bool suspended; // member offset: 72, size: 1, alignment: 1 // 3 byte(s) padding Capabilities capabilities; // member offset: 76, size: 4, alignment: 4 QTimer* speedTimer; // member offset: 80, size: 8, alignment: 8 bool isAutoDelete; // member offset: 88, size: 1, alignment: 1 // 7 byte(s) padding QEventLoop* eventLoop; // member offset: 96, size: 8, alignment: 8 QEventLoopLocker eventLoopLocker; // member offset: 104, size: 8, alignment: 8 bool isFinished; // member offset: 112, size: 1, alignment: 1 // 7 byte(s) padding }; // size: 120, alignment: 8 Struct KPluginLoaderPrivate (DW_TAG_class_type, offset 1584914) is sub-optimally packed: 37/48 bytes, 296/384 bits, optimal size is: 40 bytes class KPluginLoaderPrivate // location: /k/kde5/src/kcoreaddons/src/lib/plugin/kpluginloader.cpp:37 { KPluginLoader* q_ptr; // member offset: 0, size: 8, alignment: 8 QString const name; // member offset: 8, size: 8, alignment: 8 quint32 pluginVersion; // member offset: 16, size: 4, alignment: 4 // 4 byte(s) padding
Re: Review Request 122459: sync the bookmarkmanager with the shared one only if change was from this process
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122459/ --- (Updated Feb. 9, 2015, 5:22 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks. Bugs: 343735 https://bugs.kde.org/show_bug.cgi?id=343735 Repository: kio Description --- This is an attempt to fix a bug that seems quite difficult to reproduce, but definitely happens It seems there is a weird race condition linked to the sync of the two files .local/share/user-places.xbel .local/share/kfileplaces/bookmarks.xml causing the bug https://bugs.kde.org/show_bug.cgi?id=343735 all processes that are listening for changes in the bookmark manager will try to sync their bookmarks with the shared ones (user-places.xbel) and depending what process arrives first, some bookmarks may be lost if this is ok, it will have to go in kdelibs4 (since distributions will support kde4 apps for a long time still) as well as (ouch) dolphin both 4 and 5, since it forks KFilePlacesModel Diffs - src/filewidgets/kfileplacesmodel.cpp d8da868 src/filewidgets/kfileplacessharedbookmarks.cpp fbd6b53 src/filewidgets/kfileplacessharedbookmarks_p.h 4c9f6a4 Diff: https://git.reviewboard.kde.org/r/122459/diff/ Testing --- Thanks, Marco Martin ___ 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/#review75723 --- Ship it! Ship It! - Marco Martin On Feb. 9, 2015, 3:31 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, 3:31 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/#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: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated feb. 9, 2015, 5:51 p.m.) Review request for KDE Frameworks. Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing (updated) --- Besides tests, tried smb kioslave and it worked great. Thanks, Àlex Fiestas ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 122478: KConfigWidgets: Remove leftover uses of tr() for translation
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122478/ --- (Updated Feb. 9, 2015, 5:29 p.m.) Status -- This change has been marked as submitted. Review request for KDE Frameworks. Repository: kconfigwidgets Description --- Remove leftover uses of tr() for translation KConfigWidgets uses KI18n and these strings were not even extracted. Just noticed this randomly today. Two tr() calls still remain in ktipdialog.cpp. Those translate the tips themselves. There is a bug report at https://bugs.kde.org/show_bug.cgi?id=343936. Diffs - src/kconfigdialog.cpp 6214348 src/ktipdialog.cpp bd3cbd3 Diff: https://git.reviewboard.kde.org/r/122478/diff/ Testing --- Strings still show up correctly in UI. Thanks, Lasse Liehu ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated feb. 9, 2015, 5:45 p.m.) Review request for KDE Frameworks. Changes --- Add tests Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs (updated) - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing --- Thanks, Àlex Fiestas ___ 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/#review75719 --- Ship it! Seems ok to me - Ivan Čukić On Feb. 9, 2015, 3:31 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, 3:31 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
Jenkins build became unstable: kservice_master_qt5 #235
See http://build.kde.org/job/kservice_master_qt5/235/changes ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 120185: Look for kdesu in the correct location
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120185/#review75682 --- Ship it! libexec/kf5 sounds sensible to me. I'd say, barring any objections, please push this as it solves a real problem. - Sebastian Kügler On Oct. 4, 2014, 3:32 p.m., Maarten De Meyer wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/120185/ --- (Updated Oct. 4, 2014, 3:32 p.m.) Review request for KDE Frameworks and David Faure. Bugs: 338755 https://bugs.kde.org/show_bug.cgi?id=338755 Repository: kio Description --- kdesu is installed in libexec/ look for it there first. I left the findExecutable search as a backup. Is looking in CMAKE_INSTALL_FULL_LIBEXECDIR correct? Or will kde-cli-tools be installed in libexec/kf5? Insert 'kdesu' at the end to show a nicer error. If we leave this part out the error is Could not launch 'root' which is somewhat correct but not as easy to figure out as Could not launch 'kdesu' Also added an unrelated QFile::decodeName() call. Diffs - autotests/krununittest.cpp b1da9aa src/core/desktopexecparser.cpp 9510697 Diff: https://git.reviewboard.kde.org/r/120185/diff/ Testing --- Created .desktop file with X-KDE-SubstituteUID=true Now I can launch it as root and when I remove kdesu I got a normal error message. Unit test no longer skips because kdesu is not found. Thanks, Maarten De Meyer ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Review Request 122498: Register with dbus to wake up KRun, and unregister then
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122498/ --- Review request for Kate and KDE Frameworks. Repository: kate Description --- This is the patch from bug 339094. As it seemes there's no response/activity to that, I just put it here for proper review (and will commit on behalf of Liu Zhe obviously) as this is really annoying bug. Following is the description by Liu: Documentation of KToolInvocation (called by KRun when click a file in dolphin) said Multi means Always start a new service, wait until the service has registered with D-Bus. But when the second text file is opened, kate just starts a QApplication and waits, without registering with DBus. In KDE 4, the second instance of kate launchs an empty KApplication to do this job, but the author said it's too bad to register a service. Maybe we can register to make KService happy and unregister at once. Another problem is that the kate icon will jump quite a few seconds after clicking the second file. I thought the reason was KStartupInfo::appStarted() not called as it requires Q_WS_X11 macro, which is not defined in Qt 5. But the bug still exists after I fixed this. Diffs - CMakeLists.txt 408f502 config.h.cmake d15b78e kate/src/kateapp.cpp 4c26f33 kate/src/main.cpp 8852754 Diff: https://git.reviewboard.kde.org/r/122498/diff/ Testing --- Hrvoje confirmed in the bug report that this indeed fixes the issue. Thanks, Martin Klapetek ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
On feb. 9, 2015, 7:24 p.m., David Faure wrote: This should not be done if the slave has provided a UDS_MIME_TYPE. E.g. kio_smb sends mimetypes that derive from inode/directory, such as application/x-smb-server and application/x-smb-workgroup. If you think this works with the current patch, please prove it with a unittest :-) I quite not get this message sorry :/ can you explain a bit more? - Àlex --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/#review75730 --- On feb. 9, 2015, 5:51 p.m., Àlex Fiestas wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated feb. 9, 2015, 5:51 p.m.) Review request for KDE Frameworks. Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing --- Besides tests, tried smb kioslave and it worked great. Thanks, Àlex Fiestas ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Review Request 122499: include applications-menu-merged too
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/122499/ --- Review request for KDE Frameworks and David Faure. Bugs: 213972 https://bugs.kde.org/show_bug.cgi?id=213972 Repository: kservice Description --- faAs per, https://developer.gnome.org/menu-spec/ The default merge directories included in the DefaultMergeDirs element. By convention, third parties may add new Menu files in this location to create their own sub-menus. Note that a system that uses either gnome-applications.menu or kde-applications.menu depending on the desktop environment in use must still use applications-merged as the default merge directory in both cases. Implementations may chose to use .menu files with names other than application.menu for tasks or menus other than the main application menu. In that case the first part of the name of the default merge directory is derived from the name of the .menu file. For example in a system that uses a preferences.menu file to describe an additional menu, the default merge directories included in the DefaultMergeDirs element in the preferences.menu file would become $XDG_CONFIG_DIRS/menus/preferences-merged/ Diffs - src/kbuildsycoca/applications.menu f2d5253 Diff: https://git.reviewboard.kde.org/r/122499/diff/ Testing --- Thanks, Bhushan Shah ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Fixing and regulating certain types of search fields across KF5 apps
Hi, we currently have numerous dynamic search fields across KDE apps that, for performance reasons, disregard input shorter than a particular QString::length(). Examples are KMail and some things in KRunner. It's a common pattern. This is very English-centric behavior; many scripts are more complex than this, in particular CJK ones. In Korean, a single character may be composed of multiple letters and form a complex syllable. Most Korean given names are two characters, many useful words are only one character. The situation in Japanese is similar - it's written using a mix of syllabaries and ideograms. The check for 3 has good intentions but breaks real use cases for these users, such as finding mail by name. Addressing this algorithmically is not all that hard. When normalizing to Unicode NFD, those complex Korean characters turn into a sequence of combining characters for the indi- vidual letters. Syllabic and ideographic code points are assigned the character class Lo in Unicode, which is info available via QChar. An NFD pass + counting Lo chars as 2 would be simple and effective enough for this purpose. The way I'd like to see this fixed is placing an API like isMinimumSearchableLength(QString) somewhere in KF5 and to start using it everywhere. Realistically developers won't find this on their own, but it means I could go around and patch up all these places without copying code around. Which framework would be a good place for this? Cheers, Eike ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121084: Rename libmolletnetwork to avoid conflict with KDE4
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121084/ --- (Updated Feb. 9, 2015, 10:48 p.m.) Status -- This change has been marked as submitted. Review request for KDE Runtime, KDE Frameworks and Plasma. Repository: kio-extras Description --- KDE-Runtime already provides libmolletnetwork, so lets rename this one to include 5 as a suffix since many apps outside of KDE still depend on KDE-Runtime. Diffs - network/ioslave/CMakeLists.txt 06a964d network/kded/CMakeLists.txt 3be676e network/network/CMakeLists.txt c0fb43e Diff: https://git.reviewboard.kde.org/r/121084/diff/ Testing --- Thanks, Armin K. ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
On feb. 9, 2015, 7:24 p.m., David Faure wrote: This should not be done if the slave has provided a UDS_MIME_TYPE. E.g. kio_smb sends mimetypes that derive from inode/directory, such as application/x-smb-server and application/x-smb-workgroup. If you think this works with the current patch, please prove it with a unittest :-) Àlex Fiestas wrote: I quite not get this message sorry :/ can you explain a bit more? Mmm I think I get it now, will check it out (write the test) tomorrow. - Àlex --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/#review75730 --- On feb. 9, 2015, 5:51 p.m., Àlex Fiestas wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated feb. 9, 2015, 5:51 p.m.) Review request for KDE Frameworks. Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing --- Besides tests, tried smb kioslave and it worked great. Thanks, Àlex Fiestas ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel
Re: Review Request 121447: Return inode/directory when isDir returns true (kfileitem)
--- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/#review75730 --- This should not be done if the slave has provided a UDS_MIME_TYPE. E.g. kio_smb sends mimetypes that derive from inode/directory, such as application/x-smb-server and application/x-smb-workgroup. If you think this works with the current patch, please prove it with a unittest :-) - David Faure On Feb. 9, 2015, 5:51 p.m., Àlex Fiestas wrote: --- This is an automatically generated e-mail. To reply, visit: https://git.reviewboard.kde.org/r/121447/ --- (Updated Feb. 9, 2015, 5:51 p.m.) Review request for KDE Frameworks. Repository: kio Description --- If we know that the item is a dir, return directly the correct mimetype for directories. More info of why this is needed at: https://git.reviewboard.kde.org/r/120909/ Diffs - autotests/kfileitemtest.h 0ee7204 autotests/kfileitemtest.cpp 59c104e src/core/kfileitem.cpp 5894226 Diff: https://git.reviewboard.kde.org/r/121447/diff/ Testing --- Besides tests, tried smb kioslave and it worked great. Thanks, Àlex Fiestas ___ Kde-frameworks-devel mailing list Kde-frameworks-devel@kde.org https://mail.kde.org/mailman/listinfo/kde-frameworks-devel