Re: Review Request 120904: [KCharSelect] [API] Add new methods to handle characters outside BMP

2014-12-06 Thread David Faure

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/120904/#review71459
---


Indeed given the QChar API I don't see a better way.


src/kcharselect.h
https://git.reviewboard.kde.org/r/120904/#comment49857

in other methods you say P==Plane, here you say P==Page, looks like a typo?



src/kcharselect.h
https://git.reviewboard.kde.org/r/120904/#comment49858

same


- David Faure


On Oct. 30, 2014, 2:02 a.m., Christoph Feck wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/120904/
 ---
 
 (Updated Oct. 30, 2014, 2:02 a.m.)
 
 
 Review request for KDE Frameworks and Daniel Laidig.
 
 
 Repository: kwidgetsaddons
 
 
 Description
 ---
 
 To be able to fix bug 142625 we need new API in KCharSelect.
 
 I am not completely sure if we should do it in KWidgetsAddons at all, because 
 the required data updates might unreasonable blow up the size of the 
 repository, but I just came up with this proposal for the API. Note that only 
 the header diff is shown for the API review, the .cpp changes can be done 
 after we settle for something.
 
 Some things to discuss:
 * uint is used by static QChar functions as the datatype, I adopted it
 * it might make sense to have a function that returns the currently selected 
 character(s) as a simple QString, with all UTF-16 handling already done. 
 Naming suggestions?
 
 
 Diffs
 -
 
   src/kcharselect.h ec93d21 
 
 Diff: https://git.reviewboard.kde.org/r/120904/diff/
 
 
 Testing
 ---
 
 No.
 
 
 Thanks,
 
 Christoph Feck
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: MultiPart HTTP requests

2014-12-06 Thread David Faure
On Friday 05 December 2014 16:14:45 Aleix Pol wrote:
 Hi all,
 Every now and then I end up having to generate MultiPart HTTP requests to
 implement some of those fancy HTTP API's to upload images.
 
 I know we have many implementation for this, namely one in kipi, then
 forked in ktp [1], one in kdevplatform reviewboard plugin, and probably
 some more. In fact I'll have to add another implementation in Kamoso as
 well.
 
 With this e-mail, I wanted to suggest to move such MPForm into KF5 (KIO?
 KCoreAddons?) so we can re-use it, and maybe hopefully not scare people
 from implementing such API's, because it helps saving some tricky code
 anyway.
 
 Aleix
 
 [1]
 http://quickgit.kde.org/?p=ktp-text-ui.gita=blobf=image-sharer%2Fmpform.h
 http://quickgit.kde.org/?p=ktp-text-ui.gita=blobf=image-sharer%2Fmpform.cp
 p

There are more indeed, I made a list at some point:

QHttpMultiPart
khtml html_formimpl.cpp
extragear/graphics/kipi-plugins/*/*mpform*
/WebCore/platform/network/FormDataBuilder.cpp

and the validator and babelfish plugins need this too.

Did you look into QHttpMultiPart? Could it be used for all this?
If not, I'd recommend adding what's missing to it.

-- 
David Faure, fa...@kde.org, http://www.davidfaure.fr
Working on KDE Frameworks 5

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 121258: Restart KDirWatch after a failed delete operation

2014-12-06 Thread David Faure

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121258/#review71460
---

Ship it!


Ship It!

- David Faure


On Nov. 26, 2014, 6:49 p.m., Arjun AK wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/121258/
 ---
 
 (Updated Nov. 26, 2014, 6:49 p.m.)
 
 
 Review request for KDE Frameworks.
 
 
 Repository: kio
 
 
 Description
 ---
 
 KDirWatch is stopped prior to a delete operation and restarted once it is 
 complete but if it fails for some reason, KDirWatch does not get restarted.
 
 Steps to reproduce
 
 
 1. `$ touch foo`
 2. `# chattr +i foo`
 3. Try to delete the file using dolphin
 4. KDirWatch is stopped and therefore the file listing does not get updated 
 until you either press F5 or go to a new url
 
 
 Diffs
 -
 
   src/core/deletejob.cpp afc0461 
 
 Diff: https://git.reviewboard.kde.org/r/121258/diff/
 
 
 Testing
 ---
 
 
 Thanks,
 
 Arjun AK
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 119607: Support for .hidden files

2014-12-06 Thread David Faure

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/119607/#review71461
---



src/core/kcoredirlister.cpp
https://git.reviewboard.kde.org/r/119607/#comment49859

split on 2 lines, each with their own comment (more modular and precise)



src/core/kcoredirlister.cpp
https://git.reviewboard.kde.org/r/119607/#comment49862

(as per latter comment, this would then have to be done with the QFileInfo 
API)

Also, did you test the case where the first item is .? localPath might 
point to the dir then, so this would look into the *parent* dir, one level too 
high.



src/core/kcoredirlister.cpp
https://git.reviewboard.kde.org/r/119607/#comment49860

this method assumes the url is local, so why not just take a local path as 
input?



src/core/kcoredirlister.cpp
https://git.reviewboard.kde.org/r/119607/#comment49863

triple lookup in the hash
 (contains, [] , [])

Use a const_iterator instead.



src/core/kcoredirlister.cpp
https://git.reviewboard.kde.org/r/119607/#comment49864

hmm shouldn't this be mtime (of the file) rather than currentDateTime? Not 
sure I can make a case where it would make a difference though. OK, maybe 
ignore this. Well it would save a system call :-)



src/core/kfileitem.h
https://git.reviewboard.kde.org/r/119607/#comment49861

yes, make private, and friend the caller


- David Faure


On Dec. 4, 2014, 9:55 p.m., Bruno Nova wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/119607/
 ---
 
 (Updated Dec. 4, 2014, 9:55 p.m.)
 
 
 Review request for KDE Frameworks and David Faure.
 
 
 Bugs: 64740 and 246260
 https://bugs.kde.org/show_bug.cgi?id=64740
 https://bugs.kde.org/show_bug.cgi?id=246260
 
 
 Repository: kio
 
 
 Description
 ---
 
 This adds support for *.hidden* files to KDE.
 
 When listing a directory, the files/folders listed in the *.hidden* file will 
 be hidden, unless the user has chosen to show hidden files.
 
 This patch was initially based on the patch provided by Mark in Bug #246260.
 
 
 Diffs
 -
 
   src/core/kcoredirlister.cpp a31d629 
   src/core/kcoredirlister_p.h dce7dbc 
   src/core/kfileitem.h bebc241 
   src/core/kfileitem.cpp 74dc069 
 
 Diff: https://git.reviewboard.kde.org/r/119607/diff/
 
 
 Testing
 ---
 
 Built and tested the patch in Project Neon.
 Dolphin was still using KDE4/Qt4 version of the library, so I only tested 
 using the desktop folder widget, and folder desktop.
 It worked correctly when I pointed it to ~ and ~/Desktop (I added 
 .hidden there).
 However, it didn't work when I pointed it to the Desktop folder (the 
 default option, not the custom location ~/Desktop).
 More testing is required.
 
 The version for KDE4/Qt4 submitted to Bug #246260 was tested in Kubuntu 
 14.04, and it worked everywhere I tested (Dolphin, open/save dialogs, folder 
 widget and detailed/tree view in Dolphin).
 It wasn't an intensive test, though.
 
 
 Thanks,
 
 Bruno Nova
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 121265: Make modified flag work for KMainWindow::setCaption(QString, bool)

2014-12-06 Thread David Faure


 On Nov. 27, 2014, 8:18 a.m., Martin Gräßlin wrote:
  weird suggestion: why not let the window manager handle it? Expose it as an 
  X11 property which gets read by the window manager/desktop shell and 
  applies a semantic way to highlight that it's modified.

This would require changes in Qt, a bit out of scope for this shorter-term 
making-it-work patch, no?


- David


---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121265/#review71000
---


On Nov. 27, 2014, 3:10 a.m., Friedrich W. H. Kossebau wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/121265/
 ---
 
 (Updated Nov. 27, 2014, 3:10 a.m.)
 
 
 Review request for KDE Frameworks.
 
 
 Repository: kxmlgui
 
 
 Description
 ---
 
 The API dox of KMainWindow::setCaption(const QString caption, bool modified) 
 claims that the flag modified specifies whether the document is modified. 
 This displays an additional sign in the title bar, usually *.
 See 
 http://api.kde.org/frameworks-api/frameworks5-apidocs/kxmlgui/html/classKMainWindow.html#aa78364d5eeb1c1f5bd333c733378741d
 
 Currently that is not true. Instead a warning is shown on the console:
 QWidgetPrivate::setWindowModified_helper: QWidget::setWindowModified: The 
 window title does not contain a '[*]' placeholder
 
 When 'KDialog::setCaption(QString, bool)' was moved to 
 'KMainWindow::setCaption(QString, bool)' for kf5, it was dumped that the old 
 code also called ' makeStandardCaption(...)' which would add a custom 
 [modified] string if the 'modified' flag was set. See 
 http://api.kde.org/4.x-api/kdelibs-apidocs/kdeui/html/kdialog_8cpp_source.html#l00475
 
 Now KMainWindow uses the 'windowModified' property of 'QWidget', and if doing 
 so, the windowtitle needs a placeholder where to put the modified marker, by 
 the form of [*]. See 
 https://qt-project.org/doc/qt-5/qwidget.html#windowTitle-prop
 
 This patch proposes to comply to the API and to add that placeholder, so the 
 method behaves as expected. It does not use [modified] flag like the 
 kde4libs variant of the code did, but by using the Qt placeholder perhaps can 
 even integrate better into the platform (not sure if there is a hook yet).
 
 
 Diffs
 -
 
   src/kmainwindow.cpp 9562318 
 
 Diff: https://git.reviewboard.kde.org/r/121265/diff/
 
 
 Testing
 ---
 
 Modified is now properly marked to Off or On in the window title, when 
 'KMainWindow::setCaption(QString,bool)' is used. Tested with Okteta.
 
 
 Thanks,
 
 Friedrich W. H. Kossebau
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 121265: Make modified flag work for KMainWindow::setCaption(QString, bool)

2014-12-06 Thread David Faure

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121265/#review71464
---

Ship it!


- David Faure


On Nov. 27, 2014, 3:10 a.m., Friedrich W. H. Kossebau wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/121265/
 ---
 
 (Updated Nov. 27, 2014, 3:10 a.m.)
 
 
 Review request for KDE Frameworks.
 
 
 Repository: kxmlgui
 
 
 Description
 ---
 
 The API dox of KMainWindow::setCaption(const QString caption, bool modified) 
 claims that the flag modified specifies whether the document is modified. 
 This displays an additional sign in the title bar, usually *.
 See 
 http://api.kde.org/frameworks-api/frameworks5-apidocs/kxmlgui/html/classKMainWindow.html#aa78364d5eeb1c1f5bd333c733378741d
 
 Currently that is not true. Instead a warning is shown on the console:
 QWidgetPrivate::setWindowModified_helper: QWidget::setWindowModified: The 
 window title does not contain a '[*]' placeholder
 
 When 'KDialog::setCaption(QString, bool)' was moved to 
 'KMainWindow::setCaption(QString, bool)' for kf5, it was dumped that the old 
 code also called ' makeStandardCaption(...)' which would add a custom 
 [modified] string if the 'modified' flag was set. See 
 http://api.kde.org/4.x-api/kdelibs-apidocs/kdeui/html/kdialog_8cpp_source.html#l00475
 
 Now KMainWindow uses the 'windowModified' property of 'QWidget', and if doing 
 so, the windowtitle needs a placeholder where to put the modified marker, by 
 the form of [*]. See 
 https://qt-project.org/doc/qt-5/qwidget.html#windowTitle-prop
 
 This patch proposes to comply to the API and to add that placeholder, so the 
 method behaves as expected. It does not use [modified] flag like the 
 kde4libs variant of the code did, but by using the Qt placeholder perhaps can 
 even integrate better into the platform (not sure if there is a hook yet).
 
 
 Diffs
 -
 
   src/kmainwindow.cpp 9562318 
 
 Diff: https://git.reviewboard.kde.org/r/121265/diff/
 
 
 Testing
 ---
 
 Modified is now properly marked to Off or On in the window title, when 
 'KMainWindow::setCaption(QString,bool)' is used. Tested with Okteta.
 
 
 Thanks,
 
 Friedrich W. H. Kossebau
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Change in kio[master]: Fix runtime warnings about listing empty URL.

2014-12-06 Thread David Faure (Code Review)
David Faure has uploaded a new change for review.

  https://gerrit.vesnicky.cesnet.cz/r/201

Change subject: Fix runtime warnings about listing empty URL.
..

Fix runtime warnings about listing empty URL.

QWARN  : KFileDialog_UnitTest::testSetFileMode(existingfiles) 
kfiledialog_unittest(7242)/default KIO::SimpleJobPrivate::simpleJobInit: 
Invalid URL:  QUrl(  )

Change-Id: Ibeed58ebd6674a14e5ee144f74dc8b4479c632fd
---
M src/filewidgets/kurlnavigator.cpp
1 file changed, 2 insertions(+), 0 deletions(-)


  git pull ssh://gerrit.vesnicky.cesnet.cz:29418/kio refs/changes/01/201/1

diff --git a/src/filewidgets/kurlnavigator.cpp 
b/src/filewidgets/kurlnavigator.cpp
index d0fd71a..0f7fc84 100644
--- a/src/filewidgets/kurlnavigator.cpp
+++ b/src/filewidgets/kurlnavigator.cpp
@@ -529,6 +529,8 @@
 void KUrlNavigator::Private::updateButtons(int startIndex)
 {
 QUrl currentUrl = q-locationUrl();
+if (!currentUrl.isValid()) // QFileDialog::setDirectory not called yet
+return;
 
 const QString path = currentUrl.toDisplayString(QUrl::PreferLocalFile);
 

-- 
To view, visit https://gerrit.vesnicky.cesnet.cz/r/201
To unsubscribe, visit https://gerrit.vesnicky.cesnet.cz/r/settings

Gerrit-MessageType: newchange
Gerrit-Change-Id: Ibeed58ebd6674a14e5ee144f74dc8b4479c632fd
Gerrit-PatchSet: 1
Gerrit-Project: kio
Gerrit-Branch: master
Gerrit-Owner: David Faure fa...@kde.org
___
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

2014-12-06 Thread Marko Käning
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


Jenkins build is back to stable : kio_master_qt5 #449

2014-12-06 Thread KDE CI System
See http://build.kde.org/job/kio_master_qt5/449/changes

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Jenkins build is back to stable : kio_stable_qt5 #12

2014-12-06 Thread KDE CI System
See http://build.kde.org/job/kio_stable_qt5/12/changes

___
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

2014-12-06 Thread David Faure
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.

-- 
David Faure, fa...@kde.org, http://www.davidfaure.fr
Working on KDE Frameworks 5

___
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

2014-12-06 Thread Marko Käning
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

2014-12-06 Thread Marko Käning
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 

KF 5.5.0 changelog

2014-12-06 Thread David Faure
Here's the changelog I wrote for 5.5.0.

Please everyone remember to use CHANGELOG in your commits, with 
a standalone description of the issue (i.e. which doesn't use the modified 
filenames as implicit context).

I tried to grab what I could from the commit logs, but I don't always 
understand everything in e.g. plasma-framework.

-- 
David Faure, fa...@kde.org, http://www.davidfaure.fr
Working on KDE Frameworks 5
### Attica

* Use all of QT_PLUGIN_PATH paths rather than just QLibraryInfo path to look for plugins

### KActivities

* Fix plugin loading with KDE_INSTALL_USE_QT_SYS_PATHS ON

### KCMUtils

* Restore KStandardGuiItems to get consistent icons and tooltips

### KCodecs

* Introduce KEmailAddress class for email validation
* Use more robust implementation of MIME codecs taken from the KMime library
* Add KCodecs::encodeRFC2047String()

### KCompletion

* Fix PageUp/Down actions in the completion popup box

### KCoreAddons

* Add KTextToHTML class for plaintext-HTML conversion
* Add KPluginMetaData::metaDataSource()
* Allow to read KPluginMetaData from .desktop files
* Kdelibs4Migration now gives priority to distro-provided KDE4_DEFAULT_HOME

### KDeclarative

* Use Qt's method of blocking for component completion rather than our own
* Make it possible to delay initialization of object incubated from QmlObject
* Add guard when trying to access root object before component is complete

### KEmoticons

* Add KEmoticonsIntegrationPlugin for KTextToHTML from KCoreAddons

### KHTML

* A number of forward-ported fixes from kdelibs, no API changes.

### KIO

* Fix Size columns being empty in the KFileWidget detailed views
* Do not drop ASN passed to KRun when executing desktop files
* Fix passing of DESKTOP_STARTUP_ID to child process in kioexec
* Fix compilation with Qt 5.2, which also fixes a race condition
* KFileItem: cleanup overlay icon usage
* Implement back/forward side mouse buttons to navigate in the history
* Allow user to cancel out of the certificate accept duration dialog box.

### KJobWidgets

* Fix compilation with Qt 5.2.0

### KNewStuff

* Also allow absolute filepaths for configfile parameter.
* Fix compilation on Windows

### KNotifications

* Make KNotificationPlugin a public class
* KPassivePopup - Set default hide delay

### KRunner

* Add a simple cli tool to run a query on all runners

### KService

* Fix KPluginTrader::query() for old JSON
* Deprecate kservice_desktop_to_json for kcoreaddons_desktop_to_json
* Implement KPluginTrader::query() using KPluginLoader::findPlugins()
* Fix KPluginInfo::entryPath() being empty when not loaded from .desktop

### KTextEditor

* Fix bug #340212: incorrect soft-tabs alignment after beginning-of-line
* Add libgit2 compile-time check for threads support

### KWidgetsAddons

* Add class KSplitterCollapserButton, a button which appears on the side of
  a splitter handle and allows easy collapsing of the widget on the opposite side
* Support monochrome icon themes (such as breeze)

### KWindowSystem

* Add KStartupInfo::createNewStartupIdForTimestamp
* Add support for more multimedia keys
* Add support for initial mapping state of WM_HINTS
* Drop incorrect warnings when using KXMessages without QX11Info

### Plasma frameworks

* Fix compilation with Qt 5.2.0
* Fix the platformstatus kded module
* Migrate BusyIndicator, ProgressBar to QtQuick.Controls
* Add thumbnailAvailable property to PlasmaCore.WindowThumbnail

### Solid

* Fix warning: No such signal org::freedesktop::UPower::Device...

### Extra cmake modules

* Set CMAKE_INSTALL_SYSCONFDIR to /etc when CMAKE_INSTALL_PREFIX is /usr (instead of /usr/etc)
* Enable -D_USE_MATH_DEFINES on Windows

### Frameworkintegration

* Implement standardButtonText().
* Fix restoring the view mode and sizes in the file dialog

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: KF 5.5.0 changelog

2014-12-06 Thread Jeremy Whiting
Looks good to me on the attica/knewstuff entries.

On Sat, Dec 6, 2014 at 7:59 AM, David Faure fa...@kde.org wrote:

 Here's the changelog I wrote for 5.5.0.

 Please everyone remember to use CHANGELOG in your commits, with
 a standalone description of the issue (i.e. which doesn't use the modified
 filenames as implicit context).

 I tried to grab what I could from the commit logs, but I don't always
 understand everything in e.g. plasma-framework.

 --
 David Faure, fa...@kde.org, http://www.davidfaure.fr
 Working on KDE Frameworks 5

 ___
 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 121299: Add NET::OSD window type

2014-12-06 Thread Kai Uwe Broulik

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121299/#review71479
---


Ping?

- Kai Uwe Broulik


On Dez. 1, 2014, 6:49 nachm., Kai Uwe Broulik wrote:
 
 ---
 This is an automatically generated e-mail. To reply, visit:
 https://git.reviewboard.kde.org/r/121299/
 ---
 
 (Updated Dez. 1, 2014, 6:49 nachm.)
 
 
 Review request for KDE Frameworks, kwin and Martin Gräßlin.
 
 
 Repository: kwindowsystem
 
 
 Description
 ---
 
 This adds a NET::OSD window type for the OSD (eg. volume feedback) so it can 
 be placed even ontop of active fullscreen windows in contrast to the 
 Notifications.
 
 Not sure whether a kde netwm thing is required or we could just use some 
 other method in KWin to decide placement.
 
 Also dunno what impact on ABI this has, I tried to only add enum values at 
 the end.
 
 
 Diffs
 -
 
   autotests/netwininfotestclient.cpp 16ba4b3 
   src/netwm.cpp 1ccba32 
   src/netwm_def.h 0352f33 
 
 Diff: https://git.reviewboard.kde.org/r/121299/diff/
 
 
 Testing
 ---
 
 In conjunction with Review 121300 these are now ontop of fullscreen windows.
 
 
 Thanks,
 
 Kai Uwe Broulik
 


___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Debugging failures in KIO when using custom CI environment

2014-12-06 Thread Jan Kundrát

Hi,
I'm working on a testing setup which plugs into Gerrit to add info whether 
a proposed change still builds and tests continue to pass. This is using a 
modified version of KDE's own Jenkins' CI scripts, but the environment is 
very likely much different. It runs within a sanitized environment on a 
CentOS7 VM using Qt 5.3.2 from EPEL.


I'm seeing quite a few timeouts during the execution of these tests. A full 
build log is at [1], the failures start at 2014-12-06 14:48:57,448.


This is how a backtrace from a debugger attached to one of the stuck tests 
looked like:


(gdb) bt
#0  0x7f2c668eba4d in poll () from /lib64/libc.so.6
#1  0x7f2c629c8dd4 in g_main_context_iterate.isra.22 () from 
/lib64/libglib-2.0.so.0
#2  0x7f2c629c8efc in g_main_context_iteration () from 
/lib64/libglib-2.0.so.0
#3  0x7f2c676d333b in 
QEventDispatcherGlib::processEvents(QFlagsQEventLoop::ProcessEventsFlag) 
() from /lib64/libQt5Core.so.5
#4  0x7f2c676776fb in 
QEventLoop::exec(QFlagsQEventLoop::ProcessEventsFlag) () from 
/lib64/libQt5Core.so.5
#5  0x7f2c67d7611c in KJob::exec (this=0x1aa30f0) at 
/home/turbo-hipster/git/3d0386f34adf/kcoreaddons/src/lib/jobs/kjob.cpp:188
#6  0x00407e36 in MkpathJobTest::shouldCreateOneDirectory 
(this=0x7fffc9943e80) at 
/home/turbo-hipster/git/0c50b62535fe/kio/autotests/mkpathjobtest.cpp:68
#7  0x00406888 in MkpathJobTest::qt_static_metacall 
(_o=0x7fffc9943e80, _c=QMetaObject::InvokeMetaMethod, _id=3, 
_a=0x7fffc9943440) at 
/home/turbo-hipster/jobs/0c50b62535fe/01/201/1/check/check-kf5qt5-generic-test-el7/b840c48/build/autotests/mkpathjobtest.moc:100
#8  0x7f2c676840fc in QMetaMethod::invoke(QObject*, Qt::ConnectionType, 
QGenericReturnArgument, QGenericArgument, QGenericArgument, 
QGenericArgument, QGenericArgument, QGenericArgument, QGenericArgument, 
QGenericArgument, QGenericArgument, QGenericArgument, QGenericArgument) 
const ()

  from /lib64/libQt5Core.so.5
#9  0x7f2c676885cc in QMetaObject::invokeMethod(QObject*, char const*, 
Qt::ConnectionType, QGenericReturnArgument, QGenericArgument, 
QGenericArgument, QGenericArgument, QGenericArgument, QGenericArgument, 
QGenericArgument, QGenericArgument, QGenericArgument, QGenericArgument, 
QGenericArgument) ()

  from /lib64/libQt5Core.so.5
#10 0x7f2c6881eff2 in QTest::qInvokeTestMethod(char const*, char 
const*) () from /lib64/libQt5Test.so.5
#11 0x7f2c6881fa48 in QTest::qExec(QObject*, int, char**) () from 
/lib64/libQt5Test.so.5
#12 0x004067d9 in main (argc=1, argv=0x7fffc9943fa8) at 
/home/turbo-hipster/git/0c50b62535fe/kio/autotests/mkpathjobtest.cpp:145


For comparison, the tests went much further when I simply used a `su -` 
environment to run them, see [2]. (That was also against a different 
commit, but I don't see any relevant changes in there.)


- Do you see some obvious problem in what I'm doing?

With kind regards,
Jan

[1] 
http://ci-logs.kde.flaska.net/01/201/1/check/check-kf5qt5-generic-test-el7/b840c48/shell_output.log

[2] http://ci-logs.kde.flaska.net/manual/rebuilddep/0022/shell_output.log

--
Trojitá, a fast Qt IMAP e-mail client -- http://trojita.flaska.net/
___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Change in kio[master]: Fix check for missing `kioslave` executable

2014-12-06 Thread Code Review
Jan Kundrát has uploaded a new change for review.

  https://gerrit.vesnicky.cesnet.cz/r/202

Change subject: Fix check for missing `kioslave` executable
..

Fix check for missing `kioslave` executable

The old version checked a string which is guaranteed to never be empty
for being empty. It's better to check whether that file is actually an
executable file.

Change-Id: If4c0417a5f514b4232a0b3a57ad40d49724f9cfe
---
M src/core/slave.cpp
1 file changed, 3 insertions(+), 2 deletions(-)


  git pull ssh://gerrit.vesnicky.cesnet.cz:29418/kio refs/changes/02/202/1

diff --git a/src/core/slave.cpp b/src/core/slave.cpp
index 2aed27c..57dc464 100644
--- a/src/core/slave.cpp
+++ b/src/core/slave.cpp
@@ -479,8 +479,9 @@
 //qDebug()  kioslave  ,   lib_path  ,   protocol  , 
  QString()  ,   slaveAddress;
 
 const QString kioslave = CMAKE_INSTALL_FULL_LIBEXECDIR_KF5 /kioslave;
-if (kioslave.isEmpty()) {
-error_text = i18n(Can not find 'kioslave' executable);
+QFileInfo kioslaveInfo(kioslave);
+if (!kioslaveInfo.exists() || !kioslaveInfo.isFile() || 
!kioslaveInfo.isExecutable()) {
+error_text = i18n(Can not find 'kioslave' executable at '%1', 
CMAKE_INSTALL_FULL_LIBEXECDIR_KF5);
 error = KIO::ERR_CANNOT_LAUNCH_PROCESS;
 delete slave;
 return 0;

-- 
To view, visit https://gerrit.vesnicky.cesnet.cz/r/202
To unsubscribe, visit https://gerrit.vesnicky.cesnet.cz/r/settings

Gerrit-MessageType: newchange
Gerrit-Change-Id: If4c0417a5f514b4232a0b3a57ad40d49724f9cfe
Gerrit-PatchSet: 1
Gerrit-Project: kio
Gerrit-Branch: master
Gerrit-Owner: Jan Kundrát j...@kde.org
Gerrit-Reviewer: David Faure fa...@kde.org
___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Debugging failures in KIO when using custom CI environment

2014-12-06 Thread Jan Kundrát
It seems that the problem is related to how a kioslave is launched along 
with using different installation dirs. Here's what got me to the right 
path:


13032 
execve(/home/turbo-hipster/target/el7-x86_64-gcc/kf5-qt5/frameworks/kio/inst/lib64/libexec/kf5/kioslave, 
[/home/turbo-hipster/target/el7-x..., 
/home/turbo-hipster/jobs/7f0e670..., file, , 
local:/tmp/runtime-turbo-hipster...], [/* 41 vars */]) = -1 ENOENT (No 
such file or directory)


There was a bug in error checking within kioslave, see [1] for a fixes.

The real problem with CI is that the CI environment configures the build 
with an expected install prefix being 
/home/turbo-hipster/target/el7-x86_64-gcc/kf5-qt5/frameworks/kio/inst, but 
the actual installation goes somewhere else, into a temporary directory. 
This problem is masked by Jenkins on build.kde.org because it will always 
attempt to deploy the build to that directory and rsync to some shared 
depot afterwards. The checks of _proposed_ changes within Gerrit/Zuul, 
however, do not perform the rsync step, and that means that the kioslave 
executable was there.


- I have stuff to fix in the CI environment.

With kind regards,
Jan

[1] https://gerrit.vesnicky.cesnet.cz/r/202

--
Trojitá, a fast Qt IMAP e-mail client -- http://trojita.flaska.net/
___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Change in kio[master]: Add diagnostics when KIO Connection server fails to initialize

2014-12-06 Thread Code Review
Jan Kundrát has uploaded a new change for review.

  https://gerrit.vesnicky.cesnet.cz/r/203

Change subject: Add diagnostics when KIO Connection server fails to initialize
..

Add diagnostics when KIO Connection server fails to initialize

Change-Id: I84d5d2cce7e2625d286d993e2a7c724627c3a84d
---
M src/core/connectionserver.cpp
M src/core/connectionserver.h
M src/core/slave.cpp
3 files changed, 12 insertions(+), 2 deletions(-)


  git pull ssh://gerrit.vesnicky.cesnet.cz:29418/kio refs/changes/03/203/1

diff --git a/src/core/connectionserver.cpp b/src/core/connectionserver.cpp
index d523844..e913ac6 100644
--- a/src/core/connectionserver.cpp
+++ b/src/core/connectionserver.cpp
@@ -110,3 +110,8 @@
 
 conn-d-dequeue();
 }
+
+void ConnectionServer::errorString() const
+{
+return d-backend ? d-backend-errorString : QString();
+}
diff --git a/src/core/connectionserver.h b/src/core/connectionserver.h
index cce1e81..8db48eb 100644
--- a/src/core/connectionserver.h
+++ b/src/core/connectionserver.h
@@ -65,6 +65,11 @@
 Connection *nextPendingConnection();
 void setNextPendingConnection(Connection *conn);
 
+/**
+ * Textual description of an error related to the connection, if any.
+ */
+void errorString() const;
+
 Q_SIGNALS:
 void newConnection();
 
@@ -75,4 +80,4 @@
 
 } // namespace KIO
 
-#endif
\ No newline at end of file
+#endif
diff --git a/src/core/slave.cpp b/src/core/slave.cpp
index 54dab07..23cf210 100644
--- a/src/core/slave.cpp
+++ b/src/core/slave.cpp
@@ -126,7 +126,7 @@
 m_idleSince = QDateTime();
 slaveconnserver-listenForRemote();
 if (!slaveconnserver-isListening()) {
-qWarning()  KIO Connection server not listening, could not 
connect;
+qWarning()  KIO Connection server not listening, could not 
connect. Error:  slaveconnserver-errorString();
 }
 }
 ~SlavePrivate()

-- 
To view, visit https://gerrit.vesnicky.cesnet.cz/r/203
To unsubscribe, visit https://gerrit.vesnicky.cesnet.cz/r/settings

Gerrit-MessageType: newchange
Gerrit-Change-Id: I84d5d2cce7e2625d286d993e2a7c724627c3a84d
Gerrit-PatchSet: 1
Gerrit-Project: kio
Gerrit-Branch: master
Gerrit-Owner: Jan Kundrát j...@kde.org
Gerrit-Reviewer: David Faure fa...@kde.org
Gerrit-Reviewer: Sysadmin Testing Account n...@kde.org
___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


GenericConfigLocation missing in qtpaths tool

2014-12-06 Thread Marko Käning
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: GenericConfigLocation missing in qtpaths tool

2014-12-06 Thread David Faure
On Saturday 06 December 2014 20:10:33 Marko Käning wrote:
 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?

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

-- 
David Faure, fa...@kde.org, http://www.davidfaure.fr
Working on KDE Frameworks 5

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 121075: Also export inner KBookmarkOwner::FutureBookmark

2014-12-06 Thread Andrius da Costa Ribas

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121075/
---

(Updated Dec. 6, 2014, 8:20 p.m.)


Status
--

This change has been marked as submitted.


Review request for KDE Frameworks and kdewin.


Repository: kbookmarks


Description
---

Konqueror references KBookmarkOwner::FutureBookmark, which is not properly 
exported, causing the build to fail.


Diffs
-

  src/kbookmarkowner.h 4d6fe0b 

Diff: https://git.reviewboard.kde.org/r/121075/diff/


Testing
---

It builds (MSVC2013 - 64bit), and Konqueror builds after this patch (along 
other patches I'm sending to review today). Kdebase-apps is still not very 
functional, though (missing icons and weird UI).


Thanks,

Andrius da Costa Ribas

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Review Request 121372: Differeniate between bookmarks for documents and the web

2014-12-06 Thread Matthew Dawson

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121372/
---

Review request for KDE Frameworks, Localization and Translation (l10n) and 
Matthew Dawson.


Bugs: 1164383
https://bugs.kde.org/show_bug.cgi?id=1164383


Repository: kconfig


Description
---

As requested in bug #1164383, this adds a new standard shortcut
for adding bookmarks for web pages.

BUG: 1164383

I'm not sure if this is the best approach.  If this is accpted, I'll commit a 
similar change into KConfigWidgets.


Diffs
-

  src/gui/kstandardshortcut.h 5bb07fb9f060b0d5950aed251da985ce3aa46661 
  src/gui/kstandardshortcut.cpp 84007374c2836c1d61cb9b9361bd4217aa4ddc32 

Diff: https://git.reviewboard.kde.org/r/121372/diff/


Testing
---

Compiles, verified context extracted properly into pot file.


Thanks,

Matthew Dawson

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel


Re: Review Request 121372: Differeniate between bookmarks for documents and the web

2014-12-06 Thread Matthew Dawson

---
This is an automatically generated e-mail. To reply, visit:
https://git.reviewboard.kde.org/r/121372/
---

(Updated Dec. 6, 2014, 3:49 p.m.)


Review request for KDE Frameworks, Localization and Translation (l10n) and 
Matthew Dawson.


Changes
---

Copy+paste error on bug number.


Bugs: 341279
https://bugs.kde.org/show_bug.cgi?id=341279


Repository: kconfig


Description
---

As requested in bug #1164383, this adds a new standard shortcut
for adding bookmarks for web pages.

BUG: 1164383

I'm not sure if this is the best approach.  If this is accpted, I'll commit a 
similar change into KConfigWidgets.


Diffs
-

  src/gui/kstandardshortcut.h 5bb07fb9f060b0d5950aed251da985ce3aa46661 
  src/gui/kstandardshortcut.cpp 84007374c2836c1d61cb9b9361bd4217aa4ddc32 

Diff: https://git.reviewboard.kde.org/r/121372/diff/


Testing
---

Compiles, verified context extracted properly into pot file.


Thanks,

Matthew Dawson

___
Kde-frameworks-devel mailing list
Kde-frameworks-devel@kde.org
https://mail.kde.org/mailman/listinfo/kde-frameworks-devel