From matti.p.airas at nokia.com Mon Feb 20 17:38:04 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 20 Feb 2012 18:38:04 +0200 Subject: [PySide] [ANN] List address has changed Message-ID: <4F4276EC.1010505@nokia.com> Hi list, The PySide mailing list address has now changed. The new address is: pyside at qt-project.org and the respective administrative interface can be found at: http://lists.qt-project.org/mailman/listinfo/pyside All existing regular and digest subscribers have been moved to the new list. Any subscribers with disabled delivery were omitted, but they can naturally subscribe to the new list manually. The old address for the list will remain functional for some more months and will just be forwarded to the new one. However, any possible mailing list filters will need to be updated right away. Sorry for the inconvenience. :-) The change was done because PySide is now being migrated to Qt facilities, and in the new setup it's easier to use the provided services whenever possible. Best regards, Matti Airas From matti.p.airas at nokia.com Mon Feb 20 17:42:18 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 20 Feb 2012 18:42:18 +0200 Subject: [PySide] Test message, please ignore Message-ID: <4F4277EA.3030801@nokia.com> Just verifying that the list forwading is correctly set up... Cheers, ma. From robert.kern at gmail.com Mon Feb 20 17:46:07 2012 From: robert.kern at gmail.com (Robert Kern) Date: Mon, 20 Feb 2012 16:46:07 +0000 Subject: [PySide] [ANN] List address has changed In-Reply-To: <4F4276EC.1010505__44667.1478021473$1329755901$gmane$org@nokia.com> References: <4F4276EC.1010505__44667.1478021473$1329755901$gmane$org@nokia.com> Message-ID: On 2/20/12 4:38 PM, Matti Airas wrote: > Hi list, > > The PySide mailing list address has now changed. > > The new address is: > > pyside at qt-project.org > > and the respective administrative interface can be found at: > > http://lists.qt-project.org/mailman/listinfo/pyside > > All existing regular and digest subscribers have been moved to the new > list. Any subscribers with disabled delivery were omitted, but they can > naturally subscribe to the new list manually. > > The old address for the list will remain functional for some more months > and will just be forwarded to the new one. However, any possible mailing > list filters will need to be updated right away. Sorry for the > inconvenience. :-) > > The change was done because PySide is now being migrated to Qt > facilities, and in the new setup it's easier to use the provided > services whenever possible. Is it on someone's TODO list to update the GMane record for this mailing list? -- Robert Kern "I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth." -- Umberto Eco From matti.p.airas at nokia.com Mon Feb 20 17:50:14 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 20 Feb 2012 18:50:14 +0200 Subject: [PySide] PySide mailing list & other project facilities In-Reply-To: References: <4F3E0F35.30105@nokia.com> Message-ID: <4F4279C6.3030507@nokia.com> On 20.02.2012 14:36, ext anatoly techtonik wrote: > And update mailing list address in archive. > https://groups.google.com/forum/#!forum/pyside Hmm... I don't seem to have the super cow powers to do that. I wonder who are the admins for the Google group? Cheers, ma. From matti.p.airas at nokia.com Mon Feb 20 17:53:12 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 20 Feb 2012 18:53:12 +0200 Subject: [PySide] [ANN] List address has changed In-Reply-To: References: <4F4276EC.1010505__44667.1478021473$1329755901$gmane$org@nokia.com> Message-ID: <4F427A78.90703@nokia.com> On 20.02.2012 18:46, ext Robert Kern wrote: > Is it on someone's TODO list to update the GMane record for this mailing list? Again, I don't have the permissions to do that. The mails seem to come through just fine but the stated addresses are wrong. Cheers, ma. From techtonik at gmail.com Mon Feb 20 19:02:50 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Mon, 20 Feb 2012 10:02:50 -0800 (PST) Subject: [PySide] [ANN] List address has changed In-Reply-To: <4F427A78.90703@nokia.com> References: <4F4276EC.1010505__44667.1478021473$1329755901$gmane$org@nokia.com> <4F427A78.90703@nokia.com> Message-ID: <18976809.3519.1329760970288.JavaMail.geo-discussion-forums@vbfm8> On Monday, February 20, 2012 7:53:12 PM UTC+3, Matti Airas wrote: > > On 20.02.2012 18:46, ext Robert Kern wrote: > > > Is it on someone's TODO list to update the GMane record for this mailing > list? > > Again, I don't have the permissions to do that. The mails seem to come > through just fine but the stated addresses are wrong. > Me neither. But I've updated address in Google Group (and I am posting from there). -------------- next part -------------- An HTML attachment was scrubbed... URL: From ltanure at gmail.com Mon Feb 20 22:42:41 2012 From: ltanure at gmail.com (Lucas Tanure) Date: Mon, 20 Feb 2012 19:42:41 -0200 Subject: [PySide] QGraphicsView and QGraphicsScene Message-ID: Hi, I'm trying to develop an Drag And Drop with QGraphicsView and QGraphicsScene. An image comes from another widget and I drop at the scene. When I add this item I set (at dropEvent) item.setFlags(QtGui.QGraphicsItem.ItemIsSelectable | QtGui.QGraphicsItem.ItemIsMovable) self.addItem(block) Well, this works, and I can move this item on the scene. But I have two problems. First is that the first item always goes to position (0,0), but the second one I can add at any position. Just the first that anywhere I drop goes to (0,0). This is because I don't set the size of the scene, and the first one grows the scene, so for the second works. But here comes my second problem, if I set some size to this scene I lose the auto scrollbar from the view (QGraphicsView). So If drop somewhere on the scene near of the edge or out of the view, I can't go to this position , because there isn't scrollbars. But I can drop at anywhere, and the item stays there. Please Help. *CODE* *Main Controll* self.scene = dragGraphicsScene() self.ui.blocksGraphicsView.setScene(self.scene) class dragGraphicsScene(QtGui.QGraphicsScene): def __init__(self, parent=None): super(dragGraphicsScene,self).__init__(parent) def dragEnterEvent(self, event): if event.mimeData().hasFormat('image/x-block'): event.accept() else: event.ignore() def dragMoveEvent(self, event): if event.mimeData().hasFormat('image/x-block'): #event.setDropAction(QtCore.Qt.MoveAction) event.accept() else: event.ignore() def dropEvent(self, event): if event.mimeData().hasFormat('image/x-block'): itemData = event.mimeData().data('image/x-block') dataStream = QtCore.QDataStream(itemData, QtCore.QIODevice.ReadOnly) pixmap = QtGui.QPixmap() offset = QtCore.QPoint() dataStream >> pixmap >> offset block = QtGui.QGraphicsPixmapItem(pixmap) block.setPos(event.scenePos()-offset) block.setFlags(QtGui.QGraphicsItem.ItemIsSelectable | QtGui.QGraphicsItem.ItemIsMovable) self.addItem(block) if event.source() == self: event.setDropAction(QtCore.Qt.MoveAction) event.accept() else: event.acceptProposedAction() else: event.ignore() *Widget where the drag starts:* class dragWidget(QtGui.QWidget): def __init__(self, parent=None): QtGui.QWidget.__init__(self, parent) self.setAcceptDrops(True) def dragEnterEvent(self, event): if event.mimeData().hasFormat('image/x-block'): event.accept() else: event.ignore() def mousePressEvent(self, event): child = self.childAt(event.pos()) pixmap = QtGui.QPixmap(child.pixmap()) itemData = QtCore.QByteArray() dataStream = QtCore.QDataStream(itemData, QtCore.QIODevice.WriteOnly) dataStream << pixmap << QtCore.QPoint(event.pos() - child.pos()) mimeData = QtCore.QMimeData() mimeData.setData("image/x-block", itemData) #mimeData.setImageData(pixmap) drag = QtGui.QDrag(self) drag.setMimeData(mimeData) drag.setPixmap(pixmap) drag.setHotSpot(event.pos() - child.pos()) #use exec_ instead of start. drag.exec_(QtCore.Qt.CopyAction | QtCore.Qt.MoveAction) Lucas A. Tanure Alves Skype : lucas.tanure +55 (19) 88176559 -------------- next part -------------- An HTML attachment was scrubbed... URL: From ltanure at gmail.com Mon Feb 20 22:42:41 2012 From: ltanure at gmail.com (Lucas Tanure) Date: Mon, 20 Feb 2012 19:42:41 -0200 Subject: [PySide] QGraphicsView and QGraphicsScene Message-ID: Hi, I'm trying to develop an Drag And Drop with QGraphicsView and QGraphicsScene. An image comes from another widget and I drop at the scene. When I add this item I set (at dropEvent) item.setFlags(QtGui.QGraphicsItem.ItemIsSelectable | QtGui.QGraphicsItem.ItemIsMovable) self.addItem(block) Well, this works, and I can move this item on the scene. But I have two problems. First is that the first item always goes to position (0,0), but the second one I can add at any position. Just the first that anywhere I drop goes to (0,0). This is because I don't set the size of the scene, and the first one grows the scene, so for the second works. But here comes my second problem, if I set some size to this scene I lose the auto scrollbar from the view (QGraphicsView). So If drop somewhere on the scene near of the edge or out of the view, I can't go to this position , because there isn't scrollbars. But I can drop at anywhere, and the item stays there. Please Help. *CODE* *Main Controll* self.scene = dragGraphicsScene() self.ui.blocksGraphicsView.setScene(self.scene) class dragGraphicsScene(QtGui.QGraphicsScene): def __init__(self, parent=None): super(dragGraphicsScene,self).__init__(parent) def dragEnterEvent(self, event): if event.mimeData().hasFormat('image/x-block'): event.accept() else: event.ignore() def dragMoveEvent(self, event): if event.mimeData().hasFormat('image/x-block'): #event.setDropAction(QtCore.Qt.MoveAction) event.accept() else: event.ignore() def dropEvent(self, event): if event.mimeData().hasFormat('image/x-block'): itemData = event.mimeData().data('image/x-block') dataStream = QtCore.QDataStream(itemData, QtCore.QIODevice.ReadOnly) pixmap = QtGui.QPixmap() offset = QtCore.QPoint() dataStream >> pixmap >> offset block = QtGui.QGraphicsPixmapItem(pixmap) block.setPos(event.scenePos()-offset) block.setFlags(QtGui.QGraphicsItem.ItemIsSelectable | QtGui.QGraphicsItem.ItemIsMovable) self.addItem(block) if event.source() == self: event.setDropAction(QtCore.Qt.MoveAction) event.accept() else: event.acceptProposedAction() else: event.ignore() *Widget where the drag starts:* class dragWidget(QtGui.QWidget): def __init__(self, parent=None): QtGui.QWidget.__init__(self, parent) self.setAcceptDrops(True) def dragEnterEvent(self, event): if event.mimeData().hasFormat('image/x-block'): event.accept() else: event.ignore() def mousePressEvent(self, event): child = self.childAt(event.pos()) pixmap = QtGui.QPixmap(child.pixmap()) itemData = QtCore.QByteArray() dataStream = QtCore.QDataStream(itemData, QtCore.QIODevice.WriteOnly) dataStream << pixmap << QtCore.QPoint(event.pos() - child.pos()) mimeData = QtCore.QMimeData() mimeData.setData("image/x-block", itemData) #mimeData.setImageData(pixmap) drag = QtGui.QDrag(self) drag.setMimeData(mimeData) drag.setPixmap(pixmap) drag.setHotSpot(event.pos() - child.pos()) #use exec_ instead of start. drag.exec_(QtCore.Qt.CopyAction | QtCore.Qt.MoveAction) Lucas A. Tanure Alves Skype : lucas.tanure +55 (19) 88176559 -------------- next part -------------- An HTML attachment was scrubbed... URL: From robert.kern at gmail.com Tue Feb 21 00:35:16 2012 From: robert.kern at gmail.com (Robert Kern) Date: Mon, 20 Feb 2012 23:35:16 +0000 Subject: [PySide] [ANN] List address has changed In-Reply-To: <4F427A78.90703__10594.7343669415$1329756807$gmane$org@nokia.com> References: <4F4276EC.1010505__44667.1478021473$1329755901$gmane$org@nokia.com> <4F427A78.90703__10594.7343669415$1329756807$gmane$org@nokia.com> Message-ID: On 2/20/12 4:53 PM, Matti Airas wrote: > On 20.02.2012 18:46, ext Robert Kern wrote: > >> Is it on someone's TODO list to update the GMane record for this mailing list? > > Again, I don't have the permissions to do that. The mails seem to come > through just fine but the stated addresses are wrong. You can probably submit the new list here: http://gmane.org/subscribe.php And note in the Comments box that this is the new mailing list address for gmane.comp.lib.qt.pyside . -- Robert Kern "I have come to believe that the whole world is an enigma, a harmless enigma that is made terrible by our own mad attempt to interpret it as though it had an underlying truth." -- Umberto Eco From techtonik at gmail.com Tue Feb 21 10:17:34 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Tue, 21 Feb 2012 12:17:34 +0300 Subject: [PySide] BuildScripts for Python 3 (Was: from PySide import * in Python2/3) Message-ID: Hello, It seems that there is a problem with build scripts at https://github.com/PySide/BuildScripts/blob/master/README (see below) How to fix them to building PySide for Python 3 properly? $ python3 >>> import PySide >>> dir(PySide) ['__all__', '__builtins__', '__cached__', '__doc__', '__file__', '__name__', '__package__', '__path__', '__version__', '__version_info__'] >>> PySide.__version__ '1.1.1' >>> from PySide import QtCore Traceback (most recent call last): File "", line 1, in ImportError: dynamic module does not define init function (PyInit_QtCore) -- anatoly t. On Mon, Feb 20, 2012 at 3:21 PM, anatoly techtonik wrote: > Hello, > > In Python 3 'from PySide import *' gives exception illustrated below. Is it > a bug on PySide or in Python? > > > $ python > Python 2.7.2 (default, Oct 27 2011, 01:36:46) > [GCC 4.6.1 20111003 (Red Hat 4.6.1-10)] on linux2 > Type "help", "copyright", "credits" or "license" for more information. >>>> from PySide import * >>>> > > $ python3 > Python 3.2.1 (default, Jul 11 2011, 18:55:33) > [GCC 4.6.1 20110627 (Red Hat 4.6.1-1)] on linux2 > Type "help", "copyright", "credits" or "license" for more information. >>>> from PySide import * > Traceback (most recent call last): >   File "", line 1, in > ImportError: dynamic module does not define init function (PyInit_QtCore) >>>> > > -- > anatoly t. From techtonik at gmail.com Tue Feb 21 10:23:22 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Tue, 21 Feb 2012 12:23:22 +0300 Subject: [PySide] PySide mailing list & other project facilities In-Reply-To: <4F4279C6.3030507@nokia.com> References: <4F3E0F35.30105@nokia.com> <4F4279C6.3030507@nokia.com> Message-ID: On Mon, Feb 20, 2012 at 7:50 PM, Matti Airas wrote: > On 20.02.2012 14:36, ext anatoly techtonik wrote: > >> And update mailing list address in archive. >> https://groups.google.com/forum/#!forum/pyside > > > Hmm... I don't seem to have the super cow powers to do that. I wonder who > are the admins for the Google group? Luciano and Matti now have super cow powers! -- anatoly t. From techtonik at gmail.com Tue Feb 21 11:38:59 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Tue, 21 Feb 2012 02:38:59 -0800 (PST) Subject: [PySide] Python 3 support status in wiki Message-ID: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> I wonder if the current status of Python 3 support on this page http://developer.qt.nokia.com/wiki/PySide_Python_3_Support ? -------------- next part -------------- An HTML attachment was scrubbed... URL: From matti.p.airas at nokia.com Tue Feb 21 12:11:25 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 21 Feb 2012 13:11:25 +0200 Subject: [PySide] Python 3 support status in wiki In-Reply-To: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> References: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> Message-ID: <4F437BDD.9040901@nokia.com> On 21.02.2012 12:38, ext anatoly techtonik wrote: > I wonder if the current status of Python 3 support on this page > http://developer.qt.nokia.com/wiki/PySide_Python_3_Support ? Heh, seems pretty badly outdated. But it's a wiki, anyone can update it... Cheers, ma. From th.perl at gmail.com Tue Feb 21 17:49:01 2012 From: th.perl at gmail.com (Thomas Perl) Date: Tue, 21 Feb 2012 17:49:01 +0100 Subject: [PySide] BuildScripts for Python 3 (Was: from PySide import * in Python2/3) In-Reply-To: References: Message-ID: Hi, 2012/2/21 anatoly techtonik : > It seems that there is a problem with build scripts at > https://github.com/PySide/BuildScripts/blob/master/README (see below) > How to fix them to building PySide for Python 3 properly? I've now pushed an update to the BuildScripts Git repository: https://github.com/PySide/BuildScripts This includes support for building PySide for Python 3. You have to edit the environment.sh file and uncomment one line. PySide for Python 3 will be installed in a different directory, so that you can easily switch between it and PySide for Python 2 - simply edit the environment.sh and then "source" it in your shell. See the updated README file for more detailed instructions on how to build for Python 3. Please also note that pyside-tools and mobility are not yet able to utilize/find Python 3 from the CMake scripts, I've filed two bugs: http://bugs.pyside.org/show_bug.cgi?id=1148 http://bugs.pyside.org/show_bug.cgi?id=1149 HTH :) Thomas From hi5freinds at gmail.com Wed Feb 22 16:59:19 2012 From: hi5freinds at gmail.com (Ravi ..........!) Date: Wed, 22 Feb 2012 09:59:19 -0600 Subject: [PySide] Could not get focus on any of the input widgets while working with MDI application? Message-ID: Hi!, I have a MDI application which has couple of input widgtes on MainWindow and inside Docking widgets. When ever i tried to bring up a context menu by right clicking on the QMdiSubwindow, all the input widgets in MDI losing their focus. Note : QMdiSubWindow has an external window (like a xterm) which is not a Qt widget and Context menu is coming from external widget. Is this a bug in Qt ? if not how can i solve this problem? any help is appreciated thanks in advance. -------------- next part -------------- An HTML attachment was scrubbed... URL: From betak102 at gmail.com Fri Feb 24 20:18:49 2012 From: betak102 at gmail.com (BetaK102) Date: Sat, 25 Feb 2012 02:18:49 +0700 Subject: [PySide] pyside@qt-project.org Message-ID: pyside at qt-project.org -------------- next part -------------- An HTML attachment was scrubbed... URL: From th.perl at gmail.com Sat Feb 25 14:42:34 2012 From: th.perl at gmail.com (Thomas Perl) Date: Sat, 25 Feb 2012 14:42:34 +0100 Subject: [PySide] PySide Mobility maintenance status Message-ID: Hi, According to a comment in QTMOBILITY-1990[1], there will be no new releases of Qt Mobility 1.2.x, as work goes into porting Qt Mobility to Qt 5. Still, Qt Mobility 1.2.x is currently useful on Harmattan and on the Desktop (QtMultimediaKit comes to my mind). I'm only talking about the QtMultimediaKit part of Qt Mobility, as that's the only part that I currently focus on - of course, other parts should build as well. In bug 1147[2], I've reported an issue about building PySide Mobility bindings on Mac OS X with Qt 4.8 and Qt Mobility's current Git HEAD. The patch attached to that bug fixes the build for me, but in my opinion it (obviously) needs some work to be merged - including re-enabling the other modules of Qt Mobility. I also wonder if the Qt Mobility bindings could reuse the CMake macros from PySide (cmake/Macros/PySideModules.cmake) - these macros are currently defined in a slightly different form in QtMobility/CMakeLists.txt of PySide Mobility. Can these be made generic enough and shared between both projects, so both projects benefit from the changes/improvements to the macros? Who's currently maintaining PySide Mobility, and is Desktop support (Linux, OS X, Windows) a target? I'd like to work on my patch to clean it up and get it integrated upstream. With the move to the Qt Project, I'm not sure if the PySide Bugzilla is still the place to report these bugs, or if I should do these things on qt-project.org. Any pointers would be gladly appreciated :) Thanks, Thomas [1] https://bugreports.qt-project.org/browse/QTMOBILITY-1990 [2] http://bugs.pyside.org/show_bug.cgi?id=1147 From dennis.victorovich at gmail.com Sat Feb 25 14:53:29 2012 From: dennis.victorovich at gmail.com (=?utf-8?B?ZGVubmlzLnZpY3Rvcm92aWNoQGdtYWlsLmNvbQ==?=) Date: Sat, 25 Feb 2012 17:53:29 +0400 Subject: [PySide] =?utf-8?q?=D0=9D=D0=90=3A__Debug_build_of_PySide_and_Pyt?= =?utf-8?q?hon?= Message-ID: <4f48e7c9.08c9cc0a.245b.ffffe648@mx.google.com> Ъёхцхй!ххчьььллщр!тётёт! ц ----- Reply message ----- От: "S. Champailler" Кому: Тема: [PySide] Debug build of PySide and Python Дата: вс, янв 22, 2012 20:25 Dear mailing list, In case somebody needs this, here's what I've done this afternoon... I built my own PySide debug version. Which implies building a Python debug. Since I wanted to use valgrind, well, I had to build it too... Unfortunately, this doesn't help much to figure out lost pointers (which happen when ownership is not properly handled...) Anyway, I post it here... Stefan * Valgrind (bugged on Debian => I build it myself) ./configure --prefix=/home/stefan/Projects/valgrind make make install export VALGRIND_LIB=/home/stefan/Projects/valgrind/lib/valgrind * Python debug built with : Download Python source then : make clean ./configure --without-pymalloc --with-valgrind --with-pydebug --enable-shared --prefix=/home/stefan/Projects/python-valgrind --enable-unicode=ucs4 make make install Note the --enable-unicode, somehow it was hard to figure out... I'm also thinking that --with-valgrind implies --without-pymalloc... * PySide debug built with : export PYTHONBASE=$HOME/Projects/python-valgrind export PYSIDESANDBOXPATH=$HOME/Projects/PL/pyside-valgrind # Tell cmake to use our Python debug build export PYTHON_EXEC=$PYTHONBASE/bin/python export PYTHON_LIB=$PYTHONBASE/lib export PYTHON_INC=$PYTHONBASE/include export PYTHONPATH=$PYSIDESANDBOXPATH/lib/python2.7/site-packages:$PYTHONBASE export PATH=$PYSIDESANDBOXPATH/bin:$PATH export LD_LIBRARY_PATH=$PYSIDESANDBOXPATH/lib:$PYTHONBASE/lib #I guess the following is for the build only export PKG_CONFIG_PATH=$PYSIDESANDBOXPATH/lib/pkgconfig:$PKG_CONFIG_PATH # set -DENABLE_ICECC=1 if you're using the icecream distributed compiler alias runcmake='cmake .. -DCMAKE_INSTALL_PREFIX=$PYSIDESANDBOXPATH -DCMAKE_BUILD_TYPE=Debug -DENABLE_ICECC=0' and, as explianed on the wiki, but extended for the prupose of a debug build (so I've added a few var. definitions -D): cmake .. -DCMAKE_INSTALL_PREFIX=$PYSIDESANDBOXPATH -DCMAKE_BUILD_TYPE=Debug -DENABLE_ICECC=0 -DPYTHON_EXECUTABLE:PATH=$PYTHON_EXEC -DPYTHON_INCLUDE_DIRS:PATH=$PYT\ HON_INC -DPYTHON_LIBRARY:FILEPATH=$PYTHON_LIB/libpython2.7.so && make -j4 && make install ... on all the directories of PySide ("apiextractor" "generatorrunner" "shiboken" "pyside") See the scripts proposed on : http://developer.qt.nokia.com/wiki/Building_PySide_on_Linux They were my starting point. * Running in Valgrind : /home/stefan/Projects/valgrind/bin/valgrind --tool=memcheck /home/stefan/Projects/python-valgrind/bin/python python.py * Debug with gdb : export LD_LIBRARY_PATH= # This makes gdb mad, will have to redefine it in gdb environment, see below gdb /home/stefan/Projects/python-valgrind/bin/python and once in gdb : (gdb) set environment LD_LIBRARY_PATH=/home/stefan/Projects/PL/pyside-valgrind/lib:/home/stefan/Projects/python-valgrind/lib (gdb) r python.py _______________________________________________ PySide mailing list PySide at lists.pyside.org http://lists.pyside.org/listinfo/pyside -------------- next part -------------- An HTML attachment was scrubbed... URL: From dennis.victorovich at gmail.com Sun Feb 26 14:52:09 2012 From: dennis.victorovich at gmail.com (=?utf-8?B?ZGVubmlzLnZpY3Rvcm92aWNoQGdtYWlsLmNvbQ==?=) Date: Sun, 26 Feb 2012 17:52:09 +0400 Subject: [PySide] =?utf-8?q?=D0=9D=D0=90=3A__Debug_build_of_PySide_and_Pyt?= =?utf-8?q?hon?= Message-ID: <4f4a38fc.969ccc0a.29b9.1f3a@mx.google.com> Ъёхцхй!ххчьььллщр!тётёт! ц ----- Reply message ----- От: "S. Champailler" Кому: Тема: [PySide] Debug build of PySide and Python Дата: вс, янв 22, 2012 20:25 Dear mailing list, In case somebody needs this, here's what I've done this afternoon... I built my own PySide debug version. Which implies building a Python debug. Since I wanted to use valgrind, well, I had to build it too... Unfortunately, this doesn't help much to figure out lost pointers (which happen when ownership is not properly handled...) Anyway, I post it here... Stefan * Valgrind (bugged on Debian => I build it myself) ./configure --prefix=/home/stefan/Projects/valgrind make make install export VALGRIND_LIB=/home/stefan/Projects/valgrind/lib/valgrind * Python debug built with : Download Python source then : make clean ./configure --without-pymalloc --with-valgrind --with-pydebug --enable-shared --prefix=/home/stefan/Projects/python-valgrind --enable-unicode=ucs4 make make install Note the --enable-unicode, somehow it was hard to figure out... I'm also thinking that --with-valgrind implies --without-pymalloc... * PySide debug built with : export PYTHONBASE=$HOME/Projects/python-valgrind export PYSIDESANDBOXPATH=$HOME/Projects/PL/pyside-valgrind # Tell cmake to use our Python debug build export PYTHON_EXEC=$PYTHONBASE/bin/python export PYTHON_LIB=$PYTHONBASE/lib export PYTHON_INC=$PYTHONBASE/include export PYTHONPATH=$PYSIDESANDBOXPATH/lib/python2.7/site-packages:$PYTHONBASE export PATH=$PYSIDESANDBOXPATH/bin:$PATH export LD_LIBRARY_PATH=$PYSIDESANDBOXPATH/lib:$PYTHONBASE/lib #I guess the following is for the build only export PKG_CONFIG_PATH=$PYSIDESANDBOXPATH/lib/pkgconfig:$PKG_CONFIG_PATH # set -DENABLE_ICECC=1 if you're using the icecream distributed compiler alias runcmake='cmake .. -DCMAKE_INSTALL_PREFIX=$PYSIDESANDBOXPATH -DCMAKE_BUILD_TYPE=Debug -DENABLE_ICECC=0' and, as explianed on the wiki, but extended for the prupose of a debug build (so I've added a few var. definitions -D): cmake .. -DCMAKE_INSTALL_PREFIX=$PYSIDESANDBOXPATH -DCMAKE_BUILD_TYPE=Debug -DENABLE_ICECC=0 -DPYTHON_EXECUTABLE:PATH=$PYTHON_EXEC -DPYTHON_INCLUDE_DIRS:PATH=$PYT\ HON_INC -DPYTHON_LIBRARY:FILEPATH=$PYTHON_LIB/libpython2.7.so && make -j4 && make install ... on all the directories of PySide ("apiextractor" "generatorrunner" "shiboken" "pyside") See the scripts proposed on : http://developer.qt.nokia.com/wiki/Building_PySide_on_Linux They were my starting point. * Running in Valgrind : /home/stefan/Projects/valgrind/bin/valgrind --tool=memcheck /home/stefan/Projects/python-valgrind/bin/python python.py * Debug with gdb : export LD_LIBRARY_PATH= # This makes gdb mad, will have to redefine it in gdb environment, see below gdb /home/stefan/Projects/python-valgrind/bin/python and once in gdb : (gdb) set environment LD_LIBRARY_PATH=/home/stefan/Projects/PL/pyside-valgrind/lib:/home/stefan/Projects/python-valgrind/lib (gdb) r python.py _______________________________________________ PySide mailing list PySide at lists.pyside.org http://lists.pyside.org/listinfo/pyside -------------- next part -------------- An HTML attachment was scrubbed... URL: From techtonik at gmail.com Mon Feb 27 07:44:40 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Sun, 26 Feb 2012 22:44:40 -0800 (PST) Subject: [PySide] Python 3 support status in wiki In-Reply-To: <4F437BDD.9040901@nokia.com> References: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> <4F437BDD.9040901@nokia.com> Message-ID: <18783525.2648.1330325080127.JavaMail.geo-discussion-forums@vbbed8> On Tuesday, February 21, 2012 2:11:25 PM UTC+3, Matti Airas wrote: > > On 21.02.2012 12:38, ext anatoly techtonik wrote: > > I wonder if the current status of Python 3 support on this page > > http://developer.qt.nokia.com/wiki/PySide_Python_3_Support ? > > Heh, seems pretty badly outdated. But it's a wiki, anyone can update it... > Anyone who knows implementation details. That's surely not me. =) -------------- next part -------------- An HTML attachment was scrubbed... URL: From rutsky.vladimir at gmail.com Mon Feb 27 13:47:41 2012 From: rutsky.vladimir at gmail.com (Vladimir Rutsky) Date: Mon, 27 Feb 2012 16:47:41 +0400 Subject: [PySide] Debug build of PySide with release Python Message-ID: Hello! Is it possible to build debug version of PySide (so that it will use Qt debug libraries QtCored4.dll, QtGuid4.dll, etc) with release Python distribution on Windows? If I try to compile with "-DCMAKE_BUILD_TYPE=Debug" I'm getting error on Shiboken: [...] -- Found PythonLibs: C:/Python26/libs/python26.lib (Required is at least version "2.6") CMake Error at CMakeLists.txt:75 (message): Python debug library not found. Try compile shiboken with -DCMAKE_BUILD_TYPE=Release I want to use debug Qt libraries to check that PySide-based application don't do nasty things that leads to strange behavior. I don't want to use debug Python build, because my application depends on several python packages with native libraries and as I see I need to rebuild them all against debug version of Python. Thanks in advance, Vladimir Rutsky From matti.p.airas at nokia.com Mon Feb 27 14:32:09 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 27 Feb 2012 15:32:09 +0200 Subject: [PySide] PySide Mobility maintenance status In-Reply-To: References: Message-ID: <4F4B85D9.7040507@nokia.com> Hi Thomas, My assumption at the moment is that PySide Mobility is at the moment pretty much unmaintained. It was a major part of the Harmattan Python project, but now that no new development is being done for that, I don't think anyone is currently looking after it. I wonder would you be interested in picking up the ball? :-) I've planned to migrate PySide Mobility to the Qt infrastructure as well, but have delayed that until the migration of PySide proper is complete (still unable to file new bugs in Jira against PySide). Cheers, ma. On 25.02.2012 15:42, ext Thomas Perl wrote: > Hi, > > According to a comment in QTMOBILITY-1990[1], there will be no new > releases of Qt Mobility 1.2.x, as work goes into porting Qt Mobility > to Qt 5. Still, Qt Mobility 1.2.x is currently useful on Harmattan and > on the Desktop (QtMultimediaKit comes to my mind). I'm only talking > about the QtMultimediaKit part of Qt Mobility, as that's the only part > that I currently focus on - of course, other parts should build as > well. > > In bug 1147[2], I've reported an issue about building PySide Mobility > bindings on Mac OS X with Qt 4.8 and Qt Mobility's current Git HEAD. > The patch attached to that bug fixes the build for me, but in my > opinion it (obviously) needs some work to be merged - including > re-enabling the other modules of Qt Mobility. > > I also wonder if the Qt Mobility bindings could reuse the CMake macros > from PySide (cmake/Macros/PySideModules.cmake) - these macros are > currently defined in a slightly different form in > QtMobility/CMakeLists.txt of PySide Mobility. Can these be made > generic enough and shared between both projects, so both projects > benefit from the changes/improvements to the macros? > > Who's currently maintaining PySide Mobility, and is Desktop support > (Linux, OS X, Windows) a target? I'd like to work on my patch to clean > it up and get it integrated upstream. With the move to the Qt Project, > I'm not sure if the PySide Bugzilla is still the place to report these > bugs, or if I should do these things on qt-project.org. Any pointers > would be gladly appreciated :) > > Thanks, > Thomas > > [1] https://bugreports.qt-project.org/browse/QTMOBILITY-1990 > [2] http://bugs.pyside.org/show_bug.cgi?id=1147 > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside From matti.p.airas at nokia.com Mon Feb 27 14:38:31 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Mon, 27 Feb 2012 15:38:31 +0200 Subject: [PySide] Python 3 support status in wiki In-Reply-To: <18783525.2648.1330325080127.JavaMail.geo-discussion-forums@vbbed8> References: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> <4F437BDD.9040901@nokia.com> <18783525.2648.1330325080127.JavaMail.geo-discussion-forums@vbbed8> Message-ID: <4F4B8757.6070307@nokia.com> Hi, Fixed the page for you. Cheers, ma. On 27.02.2012 08:44, ext anatoly techtonik wrote: > On Tuesday, February 21, 2012 2:11:25 PM UTC+3, Matti Airas wrote: > > On 21.02.2012 12:38, ext anatoly techtonik wrote: > > I wonder if the current status of Python 3 support on this page > > http://developer.qt.nokia.com/wiki/PySide_Python_3_Support > ? > > Heh, seems pretty badly outdated. But it's a wiki, anyone can update > it... > > Anyone who knows implementation details. That's surely not me. =) > > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside From techtonik at gmail.com Mon Feb 27 15:57:36 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Mon, 27 Feb 2012 17:57:36 +0300 Subject: [PySide] Python 3 support status in wiki In-Reply-To: <4F4B8757.6070307@nokia.com> References: <26646377.7390.1329820739120.JavaMail.geo-discussion-forums@vbkl3> <4F437BDD.9040901@nokia.com> <18783525.2648.1330325080127.JavaMail.geo-discussion-forums@vbbed8> <4F4B8757.6070307@nokia.com> Message-ID: Much better. Thanks. -- anatoly t. On Mon, Feb 27, 2012 at 4:38 PM, Matti Airas wrote: > Hi, > > Fixed the page for you. > > Cheers, > > ma. > > On 27.02.2012 08:44, ext anatoly techtonik wrote: >> On Tuesday, February 21, 2012 2:11:25 PM UTC+3, Matti Airas wrote: >> >>     On 21.02.2012 12:38, ext anatoly techtonik wrote: >>      > I wonder if the current status of Python 3 support on this page >>      > http://developer.qt.nokia.com/wiki/PySide_Python_3_Support >>     ? >> >>     Heh, seems pretty badly outdated. But it's a wiki, anyone can update >>     it... >> >> Anyone who knows implementation details. That's surely not me. =) >> >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside From andkit at gmx.de Tue Feb 28 01:10:11 2012 From: andkit at gmx.de (Andy Kittner) Date: Tue, 28 Feb 2012 01:10:11 +0100 Subject: [PySide] Reporting bugs against github repositories? Message-ID: <20120228001011.GB10319@mancubus.hell> Hi all, Most of the current work seems to happen in the PySide repositories on github. Is it useful to report bugs against those (and if yes, should I use bugs.pyside.org or create an issue on github?) Thanks, Andy -- All warranty and guarantee clauses become null and void upon payment of invoice. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From paulo.alcantara at openbossa.org Tue Feb 28 03:00:54 2012 From: paulo.alcantara at openbossa.org (paulo alcantara) Date: Mon, 27 Feb 2012 23:00:54 -0300 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: <20120228001011.GB10319@mancubus.hell> References: <20120228001011.GB10319@mancubus.hell> Message-ID: Hi Andy, On Mon, Feb 27, 2012 at 9:10 PM, Andy Kittner wrote: > Hi all, > > Most of the current work seems to happen in the PySide repositories on > github. Is it useful to report bugs against those (and if yes, should I use > bugs.pyside.org or create an issue on github?) > > As long as we're still unable to file new bugs in Jira against PySide - you shoud use bugs.pyside.org instead to file new bugs. Use github.org only to either send patches or request pulls. > Thanks, > Andy > -- > All warranty and guarantee clauses become null and void upon payment of > invoice. > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > - Paulo Alcantara -------------- next part -------------- An HTML attachment was scrubbed... URL: From matti.p.airas at nokia.com Tue Feb 28 09:51:26 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 10:51:26 +0200 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: <20120228001011.GB10319@mancubus.hell> References: <20120228001011.GB10319@mancubus.hell> Message-ID: <4F4C958E.2020601@nokia.com> On 28.02.2012 02:10, ext Andy Kittner wrote: > Hi all, > > Most of the current work seems to happen in the PySide repositories on > github. Is it useful to report bugs against those (and if yes, should I > use bugs.pyside.org or create an issue on github?) Hi, The official PySide repositories are now located in Qt Project's Gerrit site: http://codereview.qt-project.org/ The repository URLs are as follows: ssh://codereview.qt-project.org:29418/pyside/apiextractor.git ssh://codereview.qt-project.org:29418/pyside/generatorrunner.git ssh://codereview.qt-project.org:29418/pyside/pyside.git ssh://codereview.qt-project.org:29418/pyside/shiboken.git I would recommend only filing bugs against code present in those repositories. I also hope that the commits present in the github fork would be merged back to the official repos soon. The new Qt Project's Jira bugtracker is unfortunately still not functional but I hope the remaining issues to be fixed in day or two. So it might be best to hold off reporting any new bugs until that. Cheers, ma. From matti.p.airas at nokia.com Tue Feb 28 10:05:05 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 11:05:05 +0200 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: References: <20120228001011.GB10319@mancubus.hell> Message-ID: <4F4C98C1.105@nokia.com> On 28.02.2012 04:00, ext paulo alcantara wrote: > Use github.org only to either send patches or > request pulls. I recommend not to send patches to github because those patches can't be directly merged back to the Gerrit repos due to potentially missing contribution agreements. Instead, could you please merge the github repos back to Gerrit ASAP so that the official repos would be up-to-date. Cheers, ma. From techtonik at gmail.com Tue Feb 28 12:21:33 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Tue, 28 Feb 2012 14:21:33 +0300 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: <4F4C98C1.105@nokia.com> References: <20120228001011.GB10319@mancubus.hell> <4F4C98C1.105@nokia.com> Message-ID: On Tue, Feb 28, 2012 at 12:05 PM, Matti Airas wrote: > On 28.02.2012 04:00, ext paulo alcantara wrote: > >> Use github.org only to either send patches or >> request pulls. github.com Is it possible to disable issues or put a notice for users who are starting new issues? > I recommend not to send patches to github because those patches can't be > directly merged back to the Gerrit repos due to potentially missing > contribution agreements. And Jira is required to sign this agreement. Is that right? In this case is it possible to provide a separate identity server, where users could authorize and sign agreement with a GitHub login? > Instead, could you please merge the github repos back to Gerrit ASAP so > that the official repos would be up-to-date. Gerrit is closed from anonymous. That's not an open process. =/ -- anatoly t. From th.perl at gmail.com Tue Feb 28 12:31:09 2012 From: th.perl at gmail.com (Thomas Perl) Date: Tue, 28 Feb 2012 12:31:09 +0100 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: References: <20120228001011.GB10319@mancubus.hell> <4F4C98C1.105@nokia.com> Message-ID: Hi, 2012/2/28 anatoly techtonik : > On Tue, Feb 28, 2012 at 12:05 PM, Matti Airas wrote: >> Instead, could you please merge the github repos back to Gerrit ASAP so >> that the official repos would be up-to-date. > > Gerrit is closed from anonymous. That's not an open process. =/ Yes, I'd also like to propose that the Github repository should (at some point in the future when the transition is finished) be a public mirror of the state of PySide inside Gerrit (if Gerrit/Qt Project doesn't provide any open repos itself). Thanks, Thomas From matti.p.airas at nokia.com Tue Feb 28 12:43:51 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 13:43:51 +0200 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: References: <20120228001011.GB10319@mancubus.hell> <4F4C98C1.105@nokia.com> Message-ID: <4F4CBDF7.6020709@nokia.com> On 28.02.2012 13:31, ext Thomas Perl wrote: > Yes, I'd also like to propose that the Github repository should (at > some point in the future when the transition is finished) be a public > mirror of the state of PySide inside Gerrit (if Gerrit/Qt Project > doesn't provide any open repos itself). True, that would be a good way to do it. I didn't pay attention to the fact that the Gerrit repos indeed require authentication - good for Anatoly to point that out. Qt uses Gitorious to provide read-only access for the other Qt repos. However, if we are to retain the old repos for archival purposes (which we IMO need to do, at least somewhere), we need to shuffle things around a bit. Cheers, ma. From matti.p.airas at nokia.com Tue Feb 28 13:56:05 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 14:56:05 +0200 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: <4F4CBDF7.6020709@nokia.com> References: <20120228001011.GB10319@mancubus.hell> <4F4C98C1.105@nokia.com> <4F4CBDF7.6020709@nokia.com> Message-ID: <4F4CCEE5.2050101@nokia.com> On 28.02.2012 13:43, ext Matti Airas wrote: > On 28.02.2012 13:31, ext Thomas Perl wrote: >> Yes, I'd also like to propose that the Github repository should (at >> some point in the future when the transition is finished) be a public >> mirror of the state of PySide inside Gerrit (if Gerrit/Qt Project >> doesn't provide any open repos itself). > True, that would be a good way to do it. I didn't pay attention to the > fact that the Gerrit repos indeed require authentication - good for > Anatoly to point that out. > > Qt uses Gitorious to provide read-only access for the other Qt repos. > However, if we are to retain the old repos for archival purposes (which > we IMO need to do, at least somewhere), we need to shuffle things around > a bit. OK, this is what we'll do: To align with other Qt projects (and to be able to use their mirroring scripts), I'll rename the existing gitorious repos to -archive, and we'll use the original gitorious repo URLs for the mirrors. That way, the URLs remain the same for the read-only repos. Cheers, ma. From matti.p.airas at nokia.com Tue Feb 28 15:48:58 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 16:48:58 +0200 Subject: [PySide] Reporting bugs against github repositories? In-Reply-To: <4F4CCEE5.2050101@nokia.com> References: <20120228001011.GB10319@mancubus.hell> <4F4C98C1.105@nokia.com> <4F4CBDF7.6020709@nokia.com> <4F4CCEE5.2050101@nokia.com> Message-ID: <4F4CE95A.80508@nokia.com> On 28.02.2012 14:56, ext Matti Airas wrote: > OK, this is what we'll do: > > To align with other Qt projects (and to be able to use their mirroring > scripts), I'll rename the existing gitorious repos to -archive, and > we'll use the original gitorious repo URLs for the mirrors. That way, > the URLs remain the same for the read-only repos. ... and this is now done. Cheers, ma. From matti.p.airas at nokia.com Tue Feb 28 15:51:57 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 16:51:57 +0200 Subject: [PySide] Call for help: regression bugs to be fixed Message-ID: <4F4CEA0D.9080406@nokia.com> Hi list, As discussed previously, we had to remove some contributions due to missing contribution license agreements from PySide when migrating the project to Qt infrastructure. We now finally have a functional bugtracker at bugreports.qt-project.org, and I have submitted bugs against the missing features. Most of the bugfixes are really trivial and will most likely result in reimplementations identical to the original implementation. However, to have some resemblance of a cleanroom implementation, it'd be best if the fixes would be provided by people who did not participate in getting the original fixes merged - that is, by people other than the original core dev team (Paulo Alcantara will do fine, too ;-)). Please find below a list of issues that need to be fixed: https://bugreports.qt-project.org/browse/PYSIDE-2 https://bugreports.qt-project.org/browse/PYSIDE-3 https://bugreports.qt-project.org/browse/PYSIDE-4 https://bugreports.qt-project.org/browse/PYSIDE-5 https://bugreports.qt-project.org/browse/PYSIDE-6 https://bugreports.qt-project.org/browse/PYSIDE-7 So, if you are a developer with some interest or experience in Cmake and C++, have a look at the bugs above and provide patches using Gerrit, the Qt code review tool at http://codereview.qt-project.org/. The faster we get the fixes, the faster we can resume work in the new, fancy setup! ;-) Cheers, ma. From anderson.lizardo at openbossa.org Tue Feb 28 16:34:43 2012 From: anderson.lizardo at openbossa.org (Anderson Lizardo) Date: Tue, 28 Feb 2012 11:34:43 -0400 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: <4F4CEA0D.9080406@nokia.com> References: <4F4CEA0D.9080406@nokia.com> Message-ID: Hi Matti, On Tue, Feb 28, 2012 at 10:51 AM, Matti Airas wrote: > Most of the bugfixes are really trivial and will most likely result in > reimplementations identical to the original implementation. However, to > have some resemblance of a cleanroom implementation, it'd be best if the > fixes would be provided by people who did not participate in getting the > original fixes merged - that is, by people other than the original core > dev team (Paulo Alcantara will do fine, too ;-)). Maybe I can also? Or am I already "tainted" for this purpose? :) Anyway, some of them are really trivial (just read the description and follow the suggested fix, if any). Good for new contributors :) Best Regards, -- Anderson Lizardo Instituto Nokia de Tecnologia - INdT Manaus - Brazil From matti.p.airas at nokia.com Tue Feb 28 17:02:16 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Tue, 28 Feb 2012 18:02:16 +0200 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: References: <4F4CEA0D.9080406@nokia.com> Message-ID: <4F4CFA88.9040907@nokia.com> On 28.02.2012 17:34, ext Anderson Lizardo wrote: > Maybe I can also? Or am I already "tainted" for this purpose? :) I don't think your name was on any of the original bugs, so sure, be my guest. :-) > Anyway, some of them are really trivial (just read the description and > follow the suggested fix, if any). Good for new contributors :) Indeed, most of them are actually one-liners with just small syntactic changes! Cheers, ma. From andkit at gmx.de Tue Feb 28 19:40:12 2012 From: andkit at gmx.de (Andy Kittner) Date: Tue, 28 Feb 2012 19:40:12 +0100 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: <4F4CFA88.9040907@nokia.com> References: <4F4CEA0D.9080406@nokia.com> <4F4CFA88.9040907@nokia.com> Message-ID: <20120228184012.GB11734@mancubus.hell> On Tue, Feb 28, 2012 at 06:02:16PM +0200, Matti Airas wrote: >> Anyway, some of them are really trivial (just read the description and >> follow the suggested fix, if any). Good for new contributors :) > >Indeed, most of them are actually one-liners with just small syntactic >changes! I could probably also do at least some of them. Since I never worked with gerrit before I looked around a bit and found: http://wiki.qt-project.org/Gerrit_Introduction http://wiki.qt-project.org/Setting_up_Gerrit I guess they apply to PySide as well (including the mentioned commit-hooks?) If I understood everything correctly essentially I would need to: - clone the repository - set up the hooks - make the change - push the change back to the master ? Anything else I need to consider? Secondly I think some merge in the shiboken repository went wrong. The following files are in the repository: libshiboken/basewrapper.cpp.orig libshiboken/basewrapper_p.h.orig And the build fails because ParentInfo::children is not consistently used as a std::list. Using either the .orig files, or a list-iterator appears to work: diff --git a/libshiboken/basewrapper.cpp b/libshiboken/basewrapper.cpp index d60d301..b5156b9 100644 --- a/libshiboken/basewrapper.cpp +++ b/libshiboken/basewrapper.cpp @@ -115,7 +115,7 @@ static int SbkObject_traverse(PyObject* self, visitproc visit, void* arg) //Visit children Shiboken::ParentInfo* pInfo = sbkSelf->d->parentInfo; if (pInfo) { - std::set::const_iterator it = pInfo->children.begin(); + std::list::const_iterator it = pInfo->children.begin(); for(; it != pInfo->children.end(); ++it) Py_VISIT(*it); } Regards, Andy -- All's well that ends. -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: not available URL: From techtonik at gmail.com Wed Feb 29 08:59:45 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Wed, 29 Feb 2012 09:59:45 +0200 Subject: [PySide] BuildScripts for Python 3 (Was: from PySide import * in Python2/3) In-Reply-To: References: Message-ID: On Tue, Feb 21, 2012 at 7:49 PM, Thomas Perl wrote: > 2012/2/21 anatoly techtonik : >> It seems that there is a problem with build scripts at >> https://github.com/PySide/BuildScripts/blob/master/README (see below) >> How to fix them to building PySide for Python 3 properly? > > I've now pushed an update to the BuildScripts Git repository: > >   https://github.com/PySide/BuildScripts Thanks. I can even say how much time it could take to do it myself. =) > This includes support for building PySide for Python 3. You have to > edit the environment.sh file and uncomment one line. PySide for Python > 3 will be installed in a different directory, so that you can easily > switch between it and PySide for Python 2 - simply edit the > environment.sh and then "source" it in your shell. Is it possible to provide Python 2 and Python 3 bindings simultaneously? 2 and 3 peacefully coexist in Linux, so why isn't it possible to use both? Editing environment.sh every time is bothersome. To me a better interface are SKIP_PYTHON3 / SKIP_PYTHON2 options in environment.sh Ideally build scripts should also detect if Python 3 is present before attempt to compile it. > See the updated README file for more detailed instructions on how to > build for Python 3. With the default behaviour above, the amount of text there can be reduced. =) > Please also note that pyside-tools and mobility are not yet able to > utilize/find Python 3 from the CMake scripts, I've filed two bugs: > > http://bugs.pyside.org/show_bug.cgi?id=1148 > http://bugs.pyside.org/show_bug.cgi?id=1149 Unfortunately, PySide doesn't compile anymore even with Python 2. CMake Error at cmake/Modules/FindQtMobility.cmake:52 (message): Couldn't find any version of QtMobility. That's a bad default to require QtMobility installed. Ideally, build scripts should detect its presence, but for a quick fix I'd move all QtMobility stuff into separate README chapter and instead of requiring everybody to type "./build_and_install apiextractor generatorrunner shiboken pyside", provided an instructions just for those who needs QtMobility to type "./build_and_install qtmobility" -- anatoly t. From th.perl at gmail.com Wed Feb 29 09:04:37 2012 From: th.perl at gmail.com (Thomas Perl) Date: Wed, 29 Feb 2012 09:04:37 +0100 Subject: [PySide] BuildScripts for Python 3 (Was: from PySide import * in Python2/3) In-Reply-To: References: Message-ID: Hi, 2012/2/29 anatoly techtonik : > On Tue, Feb 21, 2012 at 7:49 PM, Thomas Perl wrote: >> This includes support for building PySide for Python 3. You have to >> edit the environment.sh file and uncomment one line. PySide for Python >> 3 will be installed in a different directory, so that you can easily >> switch between it and PySide for Python 2 - simply edit the >> environment.sh and then "source" it in your shell. > > Is it possible to provide Python 2 and Python 3 bindings > simultaneously? 2 and 3 peacefully coexist in Linux, so why isn't it > possible to use both? Editing environment.sh every time is bothersome. Right now, I think the approach used is okay - most users might still want to use Python 2 for various reasons. We can think about changing this later on. It will also require more testing, the current approach is more fail-safe (no conflicting environment variables). >> Please also note that pyside-tools and mobility are not yet able to >> utilize/find Python 3 from the CMake scripts, I've filed two bugs: >> >> http://bugs.pyside.org/show_bug.cgi?id=1148 >> http://bugs.pyside.org/show_bug.cgi?id=1149 > > Unfortunately, PySide doesn't compile anymore even with Python 2. > > CMake Error at cmake/Modules/FindQtMobility.cmake:52 (message): >  Couldn't find any version of QtMobility. > > That's a bad default to require QtMobility installed. Ideally, build > scripts should detect its presence, but for a quick fix I'd move all > QtMobility stuff into separate README chapter and instead of requiring > everybody to type "./build_and_install apiextractor generatorrunner > shiboken pyside", provided an instructions just for those who needs > QtMobility to type "./build_and_install qtmobility" Indeed, that's probably a better default. Changed now. Thanks, Thomas From techtonik at gmail.com Wed Feb 29 11:06:05 2012 From: techtonik at gmail.com (anatoly techtonik) Date: Wed, 29 Feb 2012 12:06:05 +0200 Subject: [PySide] BuildScripts for Python 3 (Was: from PySide import * in Python2/3) In-Reply-To: References: Message-ID: On Wed, Feb 29, 2012 at 11:04 AM, Thomas Perl wrote: > > 2012/2/29 anatoly techtonik : >> On Tue, Feb 21, 2012 at 7:49 PM, Thomas Perl wrote: >>> This includes support for building PySide for Python 3. You have to >>> edit the environment.sh file and uncomment one line. PySide for Python >>> 3 will be installed in a different directory, so that you can easily >>> switch between it and PySide for Python 2 - simply edit the >>> environment.sh and then "source" it in your shell. >> >> Is it possible to provide Python 2 and Python 3 bindings >> simultaneously? 2 and 3 peacefully coexist in Linux, so why isn't it >> possible to use both? Editing environment.sh every time is bothersome. > > Right now, I think the approach used is okay - most users might still > want to use Python 2 for various reasons. We can think about changing > this later on. It will also require more testing, the current approach > is more fail-safe (no conflicting environment variables). I already experience a rise of pure Python 3 projects for Qt, and expect even more at PyCon. New people to PySide and Qt will likely to use Python 3 interface, so it will be nice to be ready. What environment variables are conflicting? We still have a chance to raise the question of specifying extension path for both Python 2 and Python 3 at the same time before Python 3.3 Alpha is out this Sunday. >> That's a bad default to require QtMobility installed. Ideally, build >> scripts should detect its presence, but for a quick fix I'd move all >> QtMobility stuff into separate README chapter and instead of requiring >> everybody to type "./build_and_install apiextractor generatorrunner >> shiboken pyside", provided an instructions just for those who needs >> QtMobility to type "./build_and_install qtmobility" > > Indeed, that's probably a better default. Changed now. Is it possible to detect features and build them automatically if available? This way everybody will be happy. -- anatoly t. From matti.p.airas at nokia.com Wed Feb 29 11:23:40 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Wed, 29 Feb 2012 12:23:40 +0200 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: <20120228184012.GB11734@mancubus.hell> References: <4F4CEA0D.9080406@nokia.com> <4F4CFA88.9040907@nokia.com> <20120228184012.GB11734@mancubus.hell> Message-ID: <4F4DFCAC.2020200@nokia.com> On 28.02.2012 20:40, ext Andy Kittner wrote: > I could probably also do at least some of them. Great! :-) > Since I never worked with gerrit before I looked around a bit and found: > > http://wiki.qt-project.org/Gerrit_Introduction > http://wiki.qt-project.org/Setting_up_Gerrit > > I guess they apply to PySide as well (including the mentioned > commit-hooks?) Yes, these instructions apply to PySide as well. > If I understood everything correctly essentially I would need to: > - clone the repository > - set up the hooks > - make the change > - push the change back to the master > ? > Anything else I need to consider? The push needs to be for the refs/for/master as the target ref. And after this, you need to either notify some reviewers about the new contribution (using this mailing list or the IRC channel) or manually add reviewers using the Gerrit web interface. Then, the patches can be reviewed and hopefully approved. > Secondly I think some merge in the shiboken repository went wrong. > The following files are in the repository: > libshiboken/basewrapper.cpp.orig > libshiboken/basewrapper_p.h.orig OK, this is awkward. Don't know what happened but I had managed to leave these in the working set when flattening the repos. I've fixed the issue in my clone and requested the Qt admins to force push the fixed commit in. Thanks a lot for the heads up! > And the build fails because ParentInfo::children is not consistently > used as a std::list. OK, my bad as well. I only did minimal reverts to the offending commits and did not test whether the code compiles after that. On the other hand, this approach should save a few edits when reimplementing the set use. :-) Cheers, ma. From matti.p.airas at nokia.com Wed Feb 29 12:02:32 2012 From: matti.p.airas at nokia.com (Matti Airas) Date: Wed, 29 Feb 2012 13:02:32 +0200 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: <4F4DFCAC.2020200@nokia.com> References: <4F4CEA0D.9080406@nokia.com> <4F4CFA88.9040907@nokia.com> <20120228184012.GB11734@mancubus.hell> <4F4DFCAC.2020200@nokia.com> Message-ID: <4F4E05C8.7070300@nokia.com> On 29.02.2012 12:23, ext Matti Airas wrote: >> Secondly I think some merge in the shiboken repository went wrong. >> The following files are in the repository: >> libshiboken/basewrapper.cpp.orig >> libshiboken/basewrapper_p.h.orig > OK, this is awkward. Don't know what happened but I had managed to leave > these in the working set when flattening the repos. I've fixed the issue > in my clone and requested the Qt admins to force push the fixed commit in. > > Thanks a lot for the heads up! The .orig files have now been removed and the fixed repo force-pushed to gerrit. Unless you have done any major changes to your local copies, it's probably best to re-clone the repository. Sorry for the inconvenience... :-/ Cheers, ma. From hugo.lima at openbossa.org Wed Feb 29 18:58:17 2012 From: hugo.lima at openbossa.org (Hugo Parente Lima) Date: Wed, 29 Feb 2012 14:58:17 -0300 Subject: [PySide] About the next PySide release Message-ID: <201202291458.25547.hugo.lima@openbossa.org> Hi, Just raising a small issue: We have enough bug fixes to do a release but due to this long transition to qt-projects plus the guys that didn't accepted the CLA for they patches, mostly of the bugfixes are on github repository and the github repository already differs a lot from the gitorious and gerrit repositories so the question is: Should I do a release based on which repository? The most recent github with new bugs fixed or the gerit without the new fixes plus minor regressions due to people that didn't accepted the CLA? A hard choice... because if I do the release based on github I'll have a nice PySide release with plenty of bugs fixed but what to do next release? i.e. it doesn't really solves the problem. If I go for gerrit, maybe there's not even enough fixes for a release plus I don't want to spend my free time at home doing the re-work of applying patches and solving conflicts on the new repository instead of fixing new bugs. While there's not a complete answer for that I can't do a release and PySide stays in this limbo with active development but without a release :-/, for sure PySide can't be that way forever. To increase the mess yesterday I pushed commits into Shiboken to simplify PySide releases and development, ApiExtractor is now a static library and GeneratorRunner is no more, so you will only need two repositories to compile PySide, the one for the generator (Shiboken) and PySide itself. Comments, opinions, whatever about this issue is appreciated. Regards. -- Hugo Parente Lima -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 198 bytes Desc: This is a digitally signed message part. URL: From matti.p.airas at nokia.com Wed Feb 29 20:07:12 2012 From: matti.p.airas at nokia.com (matti.p.airas at nokia.com) Date: Wed, 29 Feb 2012 19:07:12 +0000 Subject: [PySide] About the next PySide release In-Reply-To: <201202291458.25547.hugo.lima@openbossa.org> References: <201202291458.25547.hugo.lima@openbossa.org> Message-ID: Hi, In my opinion, the regressions need to be fixed before a release can be made. As I see it, though, it should take almost no time to fix the regression bugs if we just get volunteers, and once they're fixed, re-applying the commits from the github fork back to the main repo should be no issue whatsoever because I expect the changes after the fixes to be minimal compared to the old repos. If the release would be made from Github, the exact history would be lost once the commits are merged back to the main repo. So, I would wait until we have received fixes for the regression bugs, then merge the Github changes back to Gerrit, and then make the release based on that codebase. This would also provide a nice incentive for getting the regression bugs fixed. :-) Cheers, ma. On Feb 29, 2012, at 19:58, ext Hugo Parente Lima wrote: > Hi, > > Just raising a small issue: We have enough bug fixes to do a release but due > to this long transition to qt-projects plus the guys that didn't accepted the > CLA for they patches, mostly of the bugfixes are on github repository and the > github repository already differs a lot from the gitorious and gerrit > repositories so the question is: > > Should I do a release based on which repository? The most recent github with > new bugs fixed or the gerit without the new fixes plus minor regressions due > to people that didn't accepted the CLA? > > A hard choice... because if I do the release based on github I'll have a nice > PySide release with plenty of bugs fixed but what to do next release? i.e. it > doesn't really solves the problem. > > If I go for gerrit, maybe there's not even enough fixes for a release plus I > don't want to spend my free time at home doing the re-work of applying patches > and solving conflicts on the new repository instead of fixing new bugs. > > While there's not a complete answer for that I can't do a release and PySide > stays in this limbo with active development but without a release :-/, for > sure PySide can't be that way forever. > > To increase the mess yesterday I pushed commits into Shiboken to simplify > PySide releases and development, ApiExtractor is now a static library and > GeneratorRunner is no more, so you will only need two repositories to compile > PySide, the one for the generator (Shiboken) and PySide itself. > > Comments, opinions, whatever about this issue is appreciated. > > Regards. > > -- > Hugo Parente Lima > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside From paulo.alcantara at openbossa.org Wed Feb 29 20:45:55 2012 From: paulo.alcantara at openbossa.org (paulo alcantara) Date: Wed, 29 Feb 2012 16:45:55 -0300 Subject: [PySide] About the next PySide release In-Reply-To: <201202291458.25547.hugo.lima@openbossa.org> References: <201202291458.25547.hugo.lima@openbossa.org> Message-ID: Hi Hugo, On Wed, Feb 29, 2012 at 2:58 PM, Hugo Parente Lima wrote: > > Hi, > > Just raising a small issue: We have enough bug fixes to do a release but due > to this long transition to qt-projects plus the guys that didn't accepted the > CLA for they patches, mostly of the bugfixes are on github repository and the > github repository already differs a lot from the gitorious and gerrit > repositories so the question is: > > Should I do a release based on which repository? The most recent github with > new bugs fixed or the gerit without the new fixes plus minor regressions due > to people that didn't accepted the CLA? > There's some people waiting for the new PySide release based on the recently bugs fixed available on the github repository - some of them are important and should be possibly part of the new release, IMO. > A hard choice... because if I do the release based on github I'll have a nice > PySide release with plenty of bugs fixed but what to do next release? i.e. it > doesn't really solves the problem. > > If I go for gerrit, maybe there's not even enough fixes for a release plus I > don't want to spend my free time at home doing the re-work of applying patches > and solving conflicts on the new repository instead of fixing new bugs. > I agree with you. > While there's not a complete answer for that I can't do a release and PySide > stays in this limbo with active development but without a release :-/, for > sure PySide can't be that way forever. > I think that a new release is kind of *urgent* currently. > To increase the mess yesterday I pushed commits into Shiboken to simplify > PySide releases and development, ApiExtractor is now a static library and > GeneratorRunner is no more, so you will only need two repositories to compile > PySide, the one for the generator (Shiboken) and PySide itself. > This will really make our life easier. Great work, Hugo! > Comments, opinions, whatever about this issue is appreciated. > > Regards. > > -- > Hugo Parente Lima > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > - Paulo Alcantara From andkit at gmx.de Wed Feb 29 21:40:36 2012 From: andkit at gmx.de (Andy Kittner) Date: Wed, 29 Feb 2012 21:40:36 +0100 Subject: [PySide] Call for help: regression bugs to be fixed In-Reply-To: <4F4DFCAC.2020200@nokia.com> References: <4F4CEA0D.9080406@nokia.com> <4F4CFA88.9040907@nokia.com> <20120228184012.GB11734@mancubus.hell> <4F4DFCAC.2020200@nokia.com> Message-ID: <20120229204036.GA32400@mancubus.hell> On Wed, Feb 29, 2012 at 12:23:40PM +0200, Matti Airas wrote: >On 28.02.2012 20:40, ext Andy Kittner wrote: >> I could probably also do at least some of them. > >Great! :-) > >> Since I never worked with gerrit before I looked around a bit and found: >> >> http://wiki.qt-project.org/Gerrit_Introduction >> http://wiki.qt-project.org/Setting_up_Gerrit >> >> I guess they apply to PySide as well (including the mentioned >> commit-hooks?) > >Yes, these instructions apply to PySide as well. > >> If I understood everything correctly essentially I would need to: >> - clone the repository >> - set up the hooks >> - make the change >> - push the change back to the master >> ? >> Anything else I need to consider? > >The push needs to be for the refs/for/master as the target ref. And >after this, you need to either notify some reviewers about the new >contribution (using this mailing list or the IRC channel) or manually >add reviewers using the Gerrit web interface. Then, the patches can be >reviewed and hopefully approved. I just gave it a try with #5, and apparently it worked: http://codereview.qt-project.org/#change,18299 Issue #4 also sounded vaguely familiar, so I gave it a shot as well ;) http://codereview.qt-project.org/#change,18300 I also looked at issue 3, but assertNotEqual does not appear in the mentioned file, and the test passes for me as well (I'm using a debug build of the 2.7 branch from the cpython mercurial repo, about a week old or so). Regards, Andy -- APL is a write-only language. I can write programs in APL, but I can't read any of them. -- Roy Keir -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: application/pgp-signature Size: 198 bytes Desc: not available URL: