From litltbear at gmail.com Tue Jul 2 22:05:00 2013 From: litltbear at gmail.com (litltbear at gmail.com) Date: Tue, 2 Jul 2013 13:05:00 -0700 Subject: [PySide] Qinputdialog Message-ID: 1. I want to get input from a user it will be an integer. I'm trying valueId, ok = QtGui.QInputDialog.getInt(self, 'Values id Dialog', 'Enter Id of the Value', minValue=0) and it always show a spin box. How do i remove the spin box. 2. When I display the int I use: item_0 = QtGui.QTreeWidgetItem(self.ValuesTree) item_0.setText(0, str(value.valueId)) Is there a way to display int's to the screen as nit's so they can be sorted/displayed in numerical order? ie 11, 12, 104 not 104, 11, 12. David From sean at seanfisk.com Wed Jul 3 08:43:56 2013 From: sean at seanfisk.com (Sean Fisk) Date: Wed, 3 Jul 2013 00:43:56 -0600 Subject: [PySide] Qinputdialog In-Reply-To: References: Message-ID: Hi David, I‘ve wanted to do this before. I don’t think it's possible to use QInputDialog.getInt() without a spin box, so constructing a custom QDialogseems like the proper solution. I've also include a solution with QInputDialog, but that doesn't restrict the user to positive integers only. Just give it a try and you'll see what I mean. I personally like the QDialogway. I wasn't exactly sure what you meant by the sorting, so I just threw the integers in a QListWidget. Unfortunately, the QListWidget sorting algorithmsorts lexicographically, not numerically. So I just used Python‘s list sort and re-added all the numbers to the list. Not the nicest solution, but it works. That shouldn’t be too hard to adapt to QTreeWidget. Let me know if I misunderstood any part of your question. ~ Sean Here is the code: #!/usr/bin/env python # Prompt the user for an integer and display it in a sorted list.# Python 2.7 code from PySide import QtGui, QtCore class MainWindow(QtGui.QMainWindow): def __init__(self, parent=None): super(MainWindow, self).__init__(parent) self._int_list = [] self.setCentralWidget(QtGui.QWidget()) self._layout = QtGui.QVBoxLayout(self.centralWidget()) self._int_list_widget = QtGui.QListWidget() self._layout.addWidget(self._int_list_widget) self._add_int_qinputdialog_button = QtGui.QPushButton( 'Add integer (QInputDialog)') self._add_int_qinputdialog_button.clicked.connect( self._add_int_qinputdialog) self._layout.addWidget(self._add_int_qinputdialog_button) self._add_int_qdialog_button = QtGui.QPushButton( 'Add integer (Custom QDialog)') self._add_int_qdialog_button.clicked.connect( self._add_int_qdialog) self._layout.addWidget(self._add_int_qdialog_button) def _complain_about_integer(self): QtGui.QMessageBox.critical( self, 'Invalid Input', 'Please enter a positive integer.') def _add_int_to_list(self, text): try: integer = int(text) except ValueError: self._complain_about_integer() return if integer < 0: # For the QInputDialog solution, since it doesn't prevent # negatives. self._complain_about_integer() return self._int_list.append(integer) self._int_list.sort() self._int_list_widget.clear() self._int_list_widget.addItems( [str(integer) for integer in self._int_list]) def _add_int_qinputdialog(self): text, ok = QtGui.QInputDialog.getText( self, 'Values id Dialog', 'Enter Id of the Value', inputMethodHints=QtCore.Qt.ImhDigitsOnly) if not ok: # User closed the dialog. return self._add_int_to_list(text) def _add_int_qdialog(self): dialog = IntDialog() accepted = dialog.exec_() if accepted == QtGui.QDialog.Rejected: # User closed the dialog. return self._add_int_to_list(dialog.int_text) class IntDialog(QtGui.QDialog): def __init__(self, parent=None): super(IntDialog, self).__init__(parent) self.setWindowTitle('Values id Dialog') self._layout = QtGui.QFormLayout(self) self._input_field = QtGui.QLineEdit() self._int_validator = QtGui.QIntValidator() self._int_validator.setBottom(0) self._input_field.setValidator(self._int_validator) self._layout.addRow('Enter Id of the Value', self._input_field) self._button_box = QtGui.QDialogButtonBox( QtGui.QDialogButtonBox.Ok | QtGui.QDialogButtonBox.Cancel) self._button_box.accepted.connect(self.accept) self._button_box.rejected.connect(self.reject) self._layout.addWidget(self._button_box) @property def int_text(self): return self._input_field.text() def main(argv): app = QtGui.QApplication(argv) w = MainWindow() w.show() w.raise_() return app.exec_() if __name__ == '__main__': import sys raise SystemExit(main(sys.argv)) On Tue, Jul 2, 2013 at 2:05 PM, wrote: 1. I want to get input from a user it will be an integer. I'm trying > valueId, ok = QtGui.QInputDialog.getInt(self, 'Values id Dialog', > 'Enter Id of the Value', minValue=0) > > and it always show a spin box. How do i remove the spin box. > > 2. When I display the int I use: > item_0 = QtGui.QTreeWidgetItem(self.ValuesTree) > item_0.setText(0, str(value.valueId)) > > Is there a way to display int's to the screen as nit's so they can > be sorted/displayed in numerical order? > > ie 11, 12, 104 > not 104, 11, 12. > > David > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- Sean Fisk -------------- next part -------------- An HTML attachment was scrubbed... URL: From erik.johansson at fido.se Wed Jul 3 09:47:12 2013 From: erik.johansson at fido.se (Erik Johansson) Date: Wed, 3 Jul 2013 09:47:12 +0200 Subject: [PySide] Qinputdialog In-Reply-To: References: Message-ID: You should take a look at the setData function. That should enable you to add int's and automatically get correct number sorting. Cheers, Erik On Wed, Jul 3, 2013 at 8:43 AM, Sean Fisk wrote: > Hi David, > > I‘ve wanted to do this before. I don’t think it's possible to use > QInputDialog.getInt() without a spin box, so constructing a custom QDialogseems like the proper solution. I've also include a solution with > QInputDialog, but that doesn't restrict the user to positive integers > only. Just give it a try and you'll see what I mean. I personally like the > QDialog way. > > I wasn't exactly sure what you meant by the sorting, so I just threw the > integers in a QListWidget. Unfortunately, the QListWidget sorting > algorithmsorts lexicographically, not numerically. So I just used Python‘s list sort > and re-added all the numbers to the list. Not the nicest solution, but it > works. That shouldn’t be too hard to adapt to QTreeWidget. > > Let me know if I misunderstood any part of your question. > > ~ Sean > > Here is the code: > > #!/usr/bin/env python > # Prompt the user for an integer and display it in a sorted list.# Python 2.7 code > from PySide import QtGui, QtCore > class MainWindow(QtGui.QMainWindow): > def __init__(self, parent=None): > super(MainWindow, self).__init__(parent) > > self._int_list = [] > > self.setCentralWidget(QtGui.QWidget()) > > self._layout = QtGui.QVBoxLayout(self.centralWidget()) > self._int_list_widget = QtGui.QListWidget() > self._layout.addWidget(self._int_list_widget) > > self._add_int_qinputdialog_button = QtGui.QPushButton( > 'Add integer (QInputDialog)') > self._add_int_qinputdialog_button.clicked.connect( > self._add_int_qinputdialog) > self._layout.addWidget(self._add_int_qinputdialog_button) > > self._add_int_qdialog_button = QtGui.QPushButton( > 'Add integer (Custom QDialog)') > self._add_int_qdialog_button.clicked.connect( > self._add_int_qdialog) > self._layout.addWidget(self._add_int_qdialog_button) > > def _complain_about_integer(self): > QtGui.QMessageBox.critical( > self, 'Invalid Input', 'Please enter a positive integer.') > > def _add_int_to_list(self, text): > try: > integer = int(text) > except ValueError: > self._complain_about_integer() > return > if integer < 0: > # For the QInputDialog solution, since it doesn't prevent > # negatives. > self._complain_about_integer() > return > self._int_list.append(integer) > self._int_list.sort() > self._int_list_widget.clear() > self._int_list_widget.addItems( > [str(integer) for integer in self._int_list]) > > def _add_int_qinputdialog(self): > text, ok = QtGui.QInputDialog.getText( > self, 'Values id Dialog', 'Enter Id of the Value', > inputMethodHints=QtCore.Qt.ImhDigitsOnly) > if not ok: > # User closed the dialog. > return > self._add_int_to_list(text) > > def _add_int_qdialog(self): > dialog = IntDialog() > accepted = dialog.exec_() > if accepted == QtGui.QDialog.Rejected: > # User closed the dialog. > return > self._add_int_to_list(dialog.int_text) > class IntDialog(QtGui.QDialog): > def __init__(self, parent=None): > super(IntDialog, self).__init__(parent) > > self.setWindowTitle('Values id Dialog') > > self._layout = QtGui.QFormLayout(self) > > self._input_field = QtGui.QLineEdit() > self._int_validator = QtGui.QIntValidator() > self._int_validator.setBottom(0) > self._input_field.setValidator(self._int_validator) > > self._layout.addRow('Enter Id of the Value', self._input_field) > > self._button_box = QtGui.QDialogButtonBox( > QtGui.QDialogButtonBox.Ok | QtGui.QDialogButtonBox.Cancel) > self._button_box.accepted.connect(self.accept) > self._button_box.rejected.connect(self.reject) > self._layout.addWidget(self._button_box) > > @property > def int_text(self): > return self._input_field.text() > def main(argv): > app = QtGui.QApplication(argv) > > w = MainWindow() > w.show() > w.raise_() > return app.exec_() > if __name__ == '__main__': > import sys > raise SystemExit(main(sys.argv)) > > On Tue, Jul 2, 2013 at 2:05 PM, wrote: > > 1. I want to get input from a user it will be an integer. I'm trying >> valueId, ok = QtGui.QInputDialog.getInt(self, 'Values id Dialog', >> 'Enter Id of the Value', minValue=0) >> >> and it always show a spin box. How do i remove the spin box. >> >> 2. When I display the int I use: >> item_0 = QtGui.QTreeWidgetItem(self.ValuesTree) >> item_0.setText(0, str(value.valueId)) >> >> Is there a way to display int's to the screen as nit's so they >> can be sorted/displayed in numerical order? >> >> ie 11, 12, 104 >> not 104, 11, 12. >> >> David >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> > -- > Sean Fisk > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From matthew.woehlke at kitware.com Wed Jul 3 16:53:25 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Wed, 03 Jul 2013 10:53:25 -0400 Subject: [PySide] Qinputdialog In-Reply-To: References: Message-ID: On 2013-07-03 02:43, Sean Fisk wrote: > Unfortunately, the QListWidget sorting algorithm sorts > lexicographically, not numerically. IIRC it sorts by item data with some 'magic' depending on the data type. At least I am pretty sure I've written lists with the built-in sorting and gotten numeric sort. Try instead of setText, setData(Qt.DisplayRole), and make sure the data you give is an int and not a string. -- Matthew From sean at seanfisk.com Wed Jul 3 17:24:04 2013 From: sean at seanfisk.com (Sean Fisk) Date: Wed, 3 Jul 2013 09:24:04 -0600 Subject: [PySide] Qinputdialog In-Reply-To: References: Message-ID: Nice, Eric and Matt; didn't know about setData(). On Wed, Jul 3, 2013 at 8:53 AM, Matthew Woehlke wrote: On 2013-07-03 02:43, Sean Fisk wrote: > > Unfortunately, the QListWidget sorting algorithm sorts > > lexicographically, not numerically. > > IIRC it sorts by item data with some 'magic' depending on the data type. > At least I am pretty sure I've written lists with the built-in sorting > and gotten numeric sort. > > Try instead of setText, setData(Qt.DisplayRole), and make sure the data > you give is an int and not a string. > > -- > Matthew > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- Sean Fisk -------------- next part -------------- An HTML attachment was scrubbed... URL: From erik.johansson at fido.se Fri Jul 5 13:00:17 2013 From: erik.johansson at fido.se (Erik Johansson) Date: Fri, 5 Jul 2013 13:00:17 +0200 Subject: [PySide] QSslSocket in Linux Message-ID: Hi again list I have been compiling Qt and PySide for use with autodesk maya in linux but can't seem to get QSslSocket and ssl support working. It works in osx and windows. Qt is build with ./configure -separate-debug-info -no-rpath -no-qt3support -no-phonon -no-phonon-backend -openssl-linked and openssl is found during build. When configuring pyside i see: -- Found OpenSSL: /usr/lib64/libssl.so and: -- Checking for QSslCertificate in QtNetwork -- found -- Checking for QSslCipher in QtNetwork -- found -- Checking for QSslConfiguration in QtNetwork -- found -- Checking for QSslError in QtNetwork -- found -- Checking for QSslKey in QtNetwork -- found -- Checking for QSslSocket in QtNetwork -- found But when import QtNetwork and printing out dir(QtNetwork) this is the result: ['QAbstractNetworkCache', 'QAbstractSocket', 'QAuthenticator', 'QFtp', 'QHostAddress', 'QHostInfo', 'QHttp', 'QHttpHeader', 'QHttpRequestHeader', 'QHttpResponseHeader', 'QIPv6Address', 'QLocalServer', 'QLocalSocket', 'QNetworkAccessManager', 'QNetworkAddressEntry', 'QNetworkCacheMetaData', 'QNetworkConfiguration', 'QNetworkConfigurationManager', 'QNetworkCookie', 'QNetworkCookieJar', 'QNetworkDiskCache', 'QNetworkInterface', 'QNetworkProxy', 'QNetworkProxyFactory', 'QNetworkProxyQuery', 'QNetworkReply', 'QNetworkRequest', 'QNetworkSession', 'QSsl', 'QTcpServer', 'QTcpSocket', 'QUdpSocket', 'QUrlInfo', '__doc__', '__file__', '__name__', '__package__'] and upon trying to visit an page with https the application crashes. If anyone got some ideas or knows if its some known bug please let me know. Cheers, Erik -------------- next part -------------- An HTML attachment was scrubbed... URL: From mjnowen at gmail.com Fri Jul 5 13:30:09 2013 From: mjnowen at gmail.com (Mike Owen) Date: Fri, 5 Jul 2013 12:30:09 +0100 Subject: [PySide] QSslSocket in Linux In-Reply-To: References: Message-ID: Hi Erik, Have you seen these pages? http://around-the-corner.typepad.com/adn/2012/10/building-qt-pyqt-pyside-for-maya-2013.html http://images.autodesk.com/adsk/files/pyqtmaya2013.pdf Mike On 5 July 2013 12:00, Erik Johansson wrote: > Hi again list > > > I have been compiling Qt and PySide for use with autodesk maya in linux > but can't seem to get QSslSocket and ssl support working. It works in osx > and windows. > > Qt is build with > ./configure -separate-debug-info -no-rpath -no-qt3support -no-phonon > -no-phonon-backend -openssl-linked > > and openssl is found during build. > > When configuring pyside i see: > -- Found OpenSSL: /usr/lib64/libssl.so > and: > -- Checking for QSslCertificate in QtNetwork -- found > -- Checking for QSslCipher in QtNetwork -- found > -- Checking for QSslConfiguration in QtNetwork -- found > -- Checking for QSslError in QtNetwork -- found > -- Checking for QSslKey in QtNetwork -- found > -- Checking for QSslSocket in QtNetwork -- found > > But when import QtNetwork and printing out dir(QtNetwork) this is the > result: > > ['QAbstractNetworkCache', 'QAbstractSocket', 'QAuthenticator', 'QFtp', > 'QHostAddress', 'QHostInfo', 'QHttp', 'QHttpHeader', 'QHttpRequestHeader', > 'QHttpResponseHeader', 'QIPv6Address', 'QLocalServer', 'QLocalSocket', > 'QNetworkAccessManager', 'QNetworkAddressEntry', 'QNetworkCacheMetaData', > 'QNetworkConfiguration', 'QNetworkConfigurationManager', 'QNetworkCookie', > 'QNetworkCookieJar', 'QNetworkDiskCache', 'QNetworkInterface', > 'QNetworkProxy', 'QNetworkProxyFactory', 'QNetworkProxyQuery', > 'QNetworkReply', 'QNetworkRequest', 'QNetworkSession', 'QSsl', > 'QTcpServer', 'QTcpSocket', 'QUdpSocket', 'QUrlInfo', '__doc__', > '__file__', '__name__', '__package__'] > > > and upon trying to visit an page with https the application crashes. > > > If anyone got some ideas or knows if its some known bug please let me know. > > > Cheers, > > Erik > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From erik.johansson at fido.se Fri Jul 5 14:36:37 2013 From: erik.johansson at fido.se (Erik Johansson) Date: Fri, 5 Jul 2013 14:36:37 +0200 Subject: [PySide] QSslSocket in Linux In-Reply-To: References: Message-ID: Yeah, it's those instructions I am following. When trying to open an https page maya segfaults with: QMetaObject::invokeMethod: No such method WebCore::SocketStreamHandlePrivate::socketSentData() Program received signal SIGSEGV, Segmentation fault. 0x00007fffe1a9acc4 in RSA_size () from /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 (gdb) bt #0 0x00007fffe1a9acc4 in RSA_size () from /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 #1 0x00007fffe1a281bd in RSA_verify () from /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 #2 0x00007fffb5842433 in ssl3_get_key_exchange () from /usr/lib64/libssl.so.0.9.8 #3 0x00007fffb5844c50 in ssl3_connect () from /usr/lib64/libssl.so.0.9.8 #4 0x00007fffea0878fb in ?? () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 #5 0x00007fffea088a03 in ?? () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 #6 0x00007fffea0830df in QSslSocket::qt_metacall(QMetaObject::Call, int, void**) () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 #7 0x00007fffed497849 in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #8 0x00007fffea0659c3 in ?? () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 #9 0x00007fffea0556c1 in ?? () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 #10 0x00007fffec7e6f0e in QApplicationPrivate::notify_helper(QObject*, QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 #11 0x00007fffec7edb70 in QApplication::notify(QObject*, QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 #12 0x00007ffff4c4bd6e in QmayaApplication::notify(QObject*, QEvent*) () from /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so #13 0x00007fffed4827c3 in QCoreApplication::notifyInternal(QObject*, QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #14 0x00007fffed4ae17b in ?? () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #15 0x0000003bd6a38f0e in g_main_context_dispatch () from /lib64/libglib-2.0.so.0 #16 0x0000003bd6a3c938 in ?? () from /lib64/libglib-2.0.so.0 #17 0x0000003bd6a3ca3a in g_main_context_iteration () from /lib64/libglib-2.0.so.0 #18 0x00007fffed4ae2c5 in QEventDispatcherGlib::processEvents(QFlags) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #19 0x00007fffec890aef in ?? () from /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 #20 0x00007fffed481b85 in QEventLoop::processEvents(QFlags) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #21 0x00007fffed481dec in QEventLoop::exec(QFlags) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #22 0x00007fffed483c69 in QCoreApplication::exec() () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 #23 0x00007ffff4c487a9 in Tapplication::start() () from /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so #24 0x000000000040eeaa in appmain() () #25 0x000000000041cf21 in main () So maybe its not pysides fault but some funky system lib. //Erik On Fri, Jul 5, 2013 at 1:30 PM, Mike Owen wrote: > Hi Erik, > Have you seen these pages? > > http://around-the-corner.typepad.com/adn/2012/10/building-qt-pyqt-pyside-for-maya-2013.html > http://images.autodesk.com/adsk/files/pyqtmaya2013.pdf > Mike > > > On 5 July 2013 12:00, Erik Johansson wrote: > >> Hi again list >> >> >> I have been compiling Qt and PySide for use with autodesk maya in linux >> but can't seem to get QSslSocket and ssl support working. It works in osx >> and windows. >> >> Qt is build with >> ./configure -separate-debug-info -no-rpath -no-qt3support -no-phonon >> -no-phonon-backend -openssl-linked >> >> and openssl is found during build. >> >> When configuring pyside i see: >> -- Found OpenSSL: /usr/lib64/libssl.so >> and: >> -- Checking for QSslCertificate in QtNetwork -- found >> -- Checking for QSslCipher in QtNetwork -- found >> -- Checking for QSslConfiguration in QtNetwork -- found >> -- Checking for QSslError in QtNetwork -- found >> -- Checking for QSslKey in QtNetwork -- found >> -- Checking for QSslSocket in QtNetwork -- found >> >> But when import QtNetwork and printing out dir(QtNetwork) this is the >> result: >> >> ['QAbstractNetworkCache', 'QAbstractSocket', 'QAuthenticator', 'QFtp', >> 'QHostAddress', 'QHostInfo', 'QHttp', 'QHttpHeader', 'QHttpRequestHeader', >> 'QHttpResponseHeader', 'QIPv6Address', 'QLocalServer', 'QLocalSocket', >> 'QNetworkAccessManager', 'QNetworkAddressEntry', 'QNetworkCacheMetaData', >> 'QNetworkConfiguration', 'QNetworkConfigurationManager', 'QNetworkCookie', >> 'QNetworkCookieJar', 'QNetworkDiskCache', 'QNetworkInterface', >> 'QNetworkProxy', 'QNetworkProxyFactory', 'QNetworkProxyQuery', >> 'QNetworkReply', 'QNetworkRequest', 'QNetworkSession', 'QSsl', >> 'QTcpServer', 'QTcpSocket', 'QUdpSocket', 'QUrlInfo', '__doc__', >> '__file__', '__name__', '__package__'] >> >> >> and upon trying to visit an page with https the application crashes. >> >> >> If anyone got some ideas or knows if its some known bug please let me >> know. >> >> >> Cheers, >> >> Erik >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From erik.johansson at fido.se Fri Jul 5 14:39:36 2013 From: erik.johansson at fido.se (Erik Johansson) Date: Fri, 5 Jul 2013 14:39:36 +0200 Subject: [PySide] QSslSocket in Linux In-Reply-To: References: Message-ID: Going to try on a qtbuild that was built with runtime instead of linked openssl and see if that helps. On Fri, Jul 5, 2013 at 2:36 PM, Erik Johansson wrote: > Yeah, it's those instructions I am following. > > When trying to open an https page maya segfaults with: > QMetaObject::invokeMethod: No such method > WebCore::SocketStreamHandlePrivate::socketSentData() > > Program received signal SIGSEGV, Segmentation fault. > 0x00007fffe1a9acc4 in RSA_size () from > /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 > (gdb) bt > #0 0x00007fffe1a9acc4 in RSA_size () from > /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 > #1 0x00007fffe1a281bd in RSA_verify () from > /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 > #2 0x00007fffb5842433 in ssl3_get_key_exchange () from > /usr/lib64/libssl.so.0.9.8 > #3 0x00007fffb5844c50 in ssl3_connect () from /usr/lib64/libssl.so.0.9.8 > #4 0x00007fffea0878fb in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 > #5 0x00007fffea088a03 in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 > #6 0x00007fffea0830df in QSslSocket::qt_metacall(QMetaObject::Call, int, > void**) () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 > #7 0x00007fffed497849 in QMetaObject::activate(QObject*, QMetaObject > const*, int, void**) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #8 0x00007fffea0659c3 in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 > #9 0x00007fffea0556c1 in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 > #10 0x00007fffec7e6f0e in QApplicationPrivate::notify_helper(QObject*, > QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 > #11 0x00007fffec7edb70 in QApplication::notify(QObject*, QEvent*) () from > /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 > #12 0x00007ffff4c4bd6e in QmayaApplication::notify(QObject*, QEvent*) () > from /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so > #13 0x00007fffed4827c3 in QCoreApplication::notifyInternal(QObject*, > QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #14 0x00007fffed4ae17b in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #15 0x0000003bd6a38f0e in g_main_context_dispatch () from > /lib64/libglib-2.0.so.0 > #16 0x0000003bd6a3c938 in ?? () from /lib64/libglib-2.0.so.0 > #17 0x0000003bd6a3ca3a in g_main_context_iteration () from > /lib64/libglib-2.0.so.0 > #18 0x00007fffed4ae2c5 in > QEventDispatcherGlib::processEvents(QFlags) > () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #19 0x00007fffec890aef in ?? () from > /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 > #20 0x00007fffed481b85 in > QEventLoop::processEvents(QFlags) () from > /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #21 0x00007fffed481dec in > QEventLoop::exec(QFlags) () from > /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #22 0x00007fffed483c69 in QCoreApplication::exec() () from > /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 > #23 0x00007ffff4c487a9 in Tapplication::start() () from > /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so > #24 0x000000000040eeaa in appmain() () > #25 0x000000000041cf21 in main () > > So maybe its not pysides fault but some funky system lib. > > //Erik > > > On Fri, Jul 5, 2013 at 1:30 PM, Mike Owen wrote: > >> Hi Erik, >> Have you seen these pages? >> >> http://around-the-corner.typepad.com/adn/2012/10/building-qt-pyqt-pyside-for-maya-2013.html >> http://images.autodesk.com/adsk/files/pyqtmaya2013.pdf >> Mike >> >> >> On 5 July 2013 12:00, Erik Johansson wrote: >> >>> Hi again list >>> >>> >>> I have been compiling Qt and PySide for use with autodesk maya in linux >>> but can't seem to get QSslSocket and ssl support working. It works in osx >>> and windows. >>> >>> Qt is build with >>> ./configure -separate-debug-info -no-rpath -no-qt3support -no-phonon >>> -no-phonon-backend -openssl-linked >>> >>> and openssl is found during build. >>> >>> When configuring pyside i see: >>> -- Found OpenSSL: /usr/lib64/libssl.so >>> and: >>> -- Checking for QSslCertificate in QtNetwork -- found >>> -- Checking for QSslCipher in QtNetwork -- found >>> -- Checking for QSslConfiguration in QtNetwork -- found >>> -- Checking for QSslError in QtNetwork -- found >>> -- Checking for QSslKey in QtNetwork -- found >>> -- Checking for QSslSocket in QtNetwork -- found >>> >>> But when import QtNetwork and printing out dir(QtNetwork) this is the >>> result: >>> >>> ['QAbstractNetworkCache', 'QAbstractSocket', 'QAuthenticator', 'QFtp', >>> 'QHostAddress', 'QHostInfo', 'QHttp', 'QHttpHeader', 'QHttpRequestHeader', >>> 'QHttpResponseHeader', 'QIPv6Address', 'QLocalServer', 'QLocalSocket', >>> 'QNetworkAccessManager', 'QNetworkAddressEntry', 'QNetworkCacheMetaData', >>> 'QNetworkConfiguration', 'QNetworkConfigurationManager', 'QNetworkCookie', >>> 'QNetworkCookieJar', 'QNetworkDiskCache', 'QNetworkInterface', >>> 'QNetworkProxy', 'QNetworkProxyFactory', 'QNetworkProxyQuery', >>> 'QNetworkReply', 'QNetworkRequest', 'QNetworkSession', 'QSsl', >>> 'QTcpServer', 'QTcpSocket', 'QUdpSocket', 'QUrlInfo', '__doc__', >>> '__file__', '__name__', '__package__'] >>> >>> >>> and upon trying to visit an page with https the application crashes. >>> >>> >>> If anyone got some ideas or knows if its some known bug please let me >>> know. >>> >>> >>> Cheers, >>> >>> Erik >>> >>> _______________________________________________ >>> PySide mailing list >>> PySide at qt-project.org >>> http://lists.qt-project.org/mailman/listinfo/pyside >>> >>> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From erik.johansson at fido.se Mon Jul 8 12:33:00 2013 From: erik.johansson at fido.se (Erik Johansson) Date: Mon, 8 Jul 2013 12:33:00 +0200 Subject: [PySide] QSslSocket in Linux In-Reply-To: References: Message-ID: Runtime ssl enabled QSslSocket in pyside on linux as well. On Fri, Jul 5, 2013 at 2:39 PM, Erik Johansson wrote: > Going to try on a qtbuild that was built with runtime instead of linked > openssl and see if that helps. > > > On Fri, Jul 5, 2013 at 2:36 PM, Erik Johansson wrote: > >> Yeah, it's those instructions I am following. >> >> When trying to open an https page maya segfaults with: >> QMetaObject::invokeMethod: No such method >> WebCore::SocketStreamHandlePrivate::socketSentData() >> >> Program received signal SIGSEGV, Segmentation fault. >> 0x00007fffe1a9acc4 in RSA_size () from >> /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 >> (gdb) bt >> #0 0x00007fffe1a9acc4 in RSA_size () from >> /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 >> #1 0x00007fffe1a281bd in RSA_verify () from >> /usr/autodesk/maya2012-x64/lib/libcrypto.so.6 >> #2 0x00007fffb5842433 in ssl3_get_key_exchange () from >> /usr/lib64/libssl.so.0.9.8 >> #3 0x00007fffb5844c50 in ssl3_connect () from /usr/lib64/libssl.so.0.9.8 >> #4 0x00007fffea0878fb in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 >> #5 0x00007fffea088a03 in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 >> #6 0x00007fffea0830df in QSslSocket::qt_metacall(QMetaObject::Call, int, >> void**) () from /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 >> #7 0x00007fffed497849 in QMetaObject::activate(QObject*, QMetaObject >> const*, int, void**) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #8 0x00007fffea0659c3 in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 >> #9 0x00007fffea0556c1 in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtNetwork.so.4 >> #10 0x00007fffec7e6f0e in QApplicationPrivate::notify_helper(QObject*, >> QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 >> #11 0x00007fffec7edb70 in QApplication::notify(QObject*, QEvent*) () from >> /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 >> #12 0x00007ffff4c4bd6e in QmayaApplication::notify(QObject*, QEvent*) () >> from /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so >> #13 0x00007fffed4827c3 in QCoreApplication::notifyInternal(QObject*, >> QEvent*) () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #14 0x00007fffed4ae17b in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #15 0x0000003bd6a38f0e in g_main_context_dispatch () from >> /lib64/libglib-2.0.so.0 >> #16 0x0000003bd6a3c938 in ?? () from /lib64/libglib-2.0.so.0 >> #17 0x0000003bd6a3ca3a in g_main_context_iteration () from >> /lib64/libglib-2.0.so.0 >> #18 0x00007fffed4ae2c5 in >> QEventDispatcherGlib::processEvents(QFlags) >> () from /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #19 0x00007fffec890aef in ?? () from >> /usr/autodesk/maya2012-x64/lib/libQtGui.so.4 >> #20 0x00007fffed481b85 in >> QEventLoop::processEvents(QFlags) () from >> /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #21 0x00007fffed481dec in >> QEventLoop::exec(QFlags) () from >> /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #22 0x00007fffed483c69 in QCoreApplication::exec() () from >> /usr/autodesk/maya2012-x64/lib/libQtCore.so.4 >> #23 0x00007ffff4c487a9 in Tapplication::start() () from >> /usr/autodesk/maya2012-x64/lib/libExtensionLayer.so >> #24 0x000000000040eeaa in appmain() () >> #25 0x000000000041cf21 in main () >> >> So maybe its not pysides fault but some funky system lib. >> >> //Erik >> >> >> On Fri, Jul 5, 2013 at 1:30 PM, Mike Owen wrote: >> >>> Hi Erik, >>> Have you seen these pages? >>> >>> http://around-the-corner.typepad.com/adn/2012/10/building-qt-pyqt-pyside-for-maya-2013.html >>> http://images.autodesk.com/adsk/files/pyqtmaya2013.pdf >>> Mike >>> >>> >>> On 5 July 2013 12:00, Erik Johansson wrote: >>> >>>> Hi again list >>>> >>>> >>>> I have been compiling Qt and PySide for use with autodesk maya in linux >>>> but can't seem to get QSslSocket and ssl support working. It works in osx >>>> and windows. >>>> >>>> Qt is build with >>>> ./configure -separate-debug-info -no-rpath -no-qt3support -no-phonon >>>> -no-phonon-backend -openssl-linked >>>> >>>> and openssl is found during build. >>>> >>>> When configuring pyside i see: >>>> -- Found OpenSSL: /usr/lib64/libssl.so >>>> and: >>>> -- Checking for QSslCertificate in QtNetwork -- found >>>> -- Checking for QSslCipher in QtNetwork -- found >>>> -- Checking for QSslConfiguration in QtNetwork -- found >>>> -- Checking for QSslError in QtNetwork -- found >>>> -- Checking for QSslKey in QtNetwork -- found >>>> -- Checking for QSslSocket in QtNetwork -- found >>>> >>>> But when import QtNetwork and printing out dir(QtNetwork) this is the >>>> result: >>>> >>>> ['QAbstractNetworkCache', 'QAbstractSocket', 'QAuthenticator', 'QFtp', >>>> 'QHostAddress', 'QHostInfo', 'QHttp', 'QHttpHeader', 'QHttpRequestHeader', >>>> 'QHttpResponseHeader', 'QIPv6Address', 'QLocalServer', 'QLocalSocket', >>>> 'QNetworkAccessManager', 'QNetworkAddressEntry', 'QNetworkCacheMetaData', >>>> 'QNetworkConfiguration', 'QNetworkConfigurationManager', 'QNetworkCookie', >>>> 'QNetworkCookieJar', 'QNetworkDiskCache', 'QNetworkInterface', >>>> 'QNetworkProxy', 'QNetworkProxyFactory', 'QNetworkProxyQuery', >>>> 'QNetworkReply', 'QNetworkRequest', 'QNetworkSession', 'QSsl', >>>> 'QTcpServer', 'QTcpSocket', 'QUdpSocket', 'QUrlInfo', '__doc__', >>>> '__file__', '__name__', '__package__'] >>>> >>>> >>>> and upon trying to visit an page with https the application crashes. >>>> >>>> >>>> If anyone got some ideas or knows if its some known bug please let me >>>> know. >>>> >>>> >>>> Cheers, >>>> >>>> Erik >>>> >>>> _______________________________________________ >>>> PySide mailing list >>>> PySide at qt-project.org >>>> http://lists.qt-project.org/mailman/listinfo/pyside >>>> >>>> >>> >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Tue Jul 9 03:30:19 2013 From: jpe at wingware.com (John Ehresman) Date: Mon, 08 Jul 2013 21:30:19 -0400 Subject: [PySide] PySide and Shiboken 1.2 released Message-ID: <51DB67AB.60204@wingware.com> I'm happy to announce that PySide and Shiboken version 1.2 have been released. PySide is a community supported LGPL wrapper for the Qt libraries and Shiboken is a C++ wrapping tool used by PySide. Version 1.2 contains numerous stability fixes, improved Windows installation via easy_install, and the shiboken module is now correctly installed. The pypi page for PySide is at http://pypi.python.org/pypi/PySide It can be installed via easy_install, pip and windows installer. The source tarballs can be found at: PySide: http://download.qt-project.org/official_releases/pyside/pyside-qt4.8+1.2.0.tar.bz2 Shiboken: http://download.qt-project.org/official_releases/pyside/shiboken-1.2.0.tar.bz2 Questions / feedback may be sent to the pyside mailing list at pyside at qt-project.org Thanks: Paulo Alcantara John Cummings Robin Dunn John Ehresman Roman Lacko Stefan Landvogt Hugo Parente Lima Sébastien Sablé Nathan Smith Matthew Woehlke -- John Ehresman From greatrgb at gmail.com Tue Jul 9 04:21:17 2013 From: greatrgb at gmail.com (Tony Barbieri) Date: Mon, 8 Jul 2013 22:21:17 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DB67AB.60204@wingware.com> References: <51DB67AB.60204@wingware.com> Message-ID: Excellent news! Are there release notes up anywhere? Thanks again for all the hard work! -tony On Mon, Jul 8, 2013 at 9:30 PM, John Ehresman wrote: > I'm happy to announce that PySide and Shiboken version 1.2 have been > released. PySide is a community supported LGPL wrapper for the Qt > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > 1.2 contains numerous stability fixes, improved Windows installation via > easy_install, and the shiboken module is now correctly installed. > > The pypi page for PySide is at http://pypi.python.org/pypi/PySide It > can be installed via easy_install, pip and windows installer. > > The source tarballs can be found at: > PySide: > > http://download.qt-project.org/official_releases/pyside/pyside-qt4.8+1.2.0.tar.bz2 > Shiboken: > > http://download.qt-project.org/official_releases/pyside/shiboken-1.2.0.tar.bz2 > > Questions / feedback may be sent to the pyside mailing list at > pyside at qt-project.org > > Thanks: > > Paulo Alcantara > John Cummings > Robin Dunn > John Ehresman > Roman Lacko > Stefan Landvogt > Hugo Parente Lima > Sébastien Sablé > Nathan Smith > Matthew Woehlke > > -- > John Ehresman > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- -tony -------------- next part -------------- An HTML attachment was scrubbed... URL: From mairas at iki.fi Tue Jul 9 07:23:49 2013 From: mairas at iki.fi (Matti Airas) Date: Tue, 9 Jul 2013 08:23:49 +0300 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DB67AB.60204@wingware.com> References: <51DB67AB.60204@wingware.com> Message-ID: On 9 Jul 2013 04:30, "John Ehresman" wrote: > > I'm happy to announce that PySide and Shiboken version 1.2 have been > released. PySide is a community supported LGPL wrapper for the Qt > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > 1.2 contains numerous stability fixes, improved Windows installation via > easy_install, and the shiboken module is now correctly installed. Wow! This is brilliant - congratulations! Cheers, ma. -------------- next part -------------- An HTML attachment was scrubbed... URL: From sean at seanfisk.com Tue Jul 9 07:57:42 2013 From: sean at seanfisk.com (Sean Fisk) Date: Mon, 8 Jul 2013 23:57:42 -0600 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DB67AB.60204@wingware.com> References: <51DB67AB.60204@wingware.com> Message-ID: This is great, thanks for effort! I hope to help on the next PySide release when I get some more time! On Mon, Jul 8, 2013 at 7:30 PM, John Ehresman wrote: > I'm happy to announce that PySide and Shiboken version 1.2 have been > released. PySide is a community supported LGPL wrapper for the Qt > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > 1.2 contains numerous stability fixes, improved Windows installation via > easy_install, and the shiboken module is now correctly installed. > > The pypi page for PySide is at http://pypi.python.org/pypi/PySide It > can be installed via easy_install, pip and windows installer. > > The source tarballs can be found at: > PySide: > > http://download.qt-project.org/official_releases/pyside/pyside-qt4.8+1.2.0.tar.bz2 > Shiboken: > > http://download.qt-project.org/official_releases/pyside/shiboken-1.2.0.tar.bz2 > > Questions / feedback may be sent to the pyside mailing list at > pyside at qt-project.org > > Thanks: > > Paulo Alcantara > John Cummings > Robin Dunn > John Ehresman > Roman Lacko > Stefan Landvogt > Hugo Parente Lima > Sébastien Sablé > Nathan Smith > Matthew Woehlke > > -- > John Ehresman > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- Sean Fisk -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Tue Jul 9 08:02:31 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Tue, 09 Jul 2013 02:02:31 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DB67AB.60204@wingware.com> References: <51DB67AB.60204@wingware.com> Message-ID: <51DBA777.5030601@gamry.com> On 7/8/2013 9:30 PM, John Ehresman wrote: > I'm happy to announce that PySide and Shiboken version 1.2 have been > released. PySide is a community supported LGPL wrapper for the Qt > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > 1.2 contains numerous stability fixes, improved Windows installation via > easy_install, and the shiboken module is now correctly installed. > > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It > can be installed via easy_install, pip and windows installer. The pypi link points to 1.1.2 and that is all that easy_install finds. Joel From backup.rlacko at gmail.com Tue Jul 9 08:27:13 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 08:27:13 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DBA777.5030601@gamry.com> References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> Message-ID: 2013/7/9 Joel B. Mohler > On 7/8/2013 9:30 PM, John Ehresman wrote: > > I'm happy to announce that PySide and Shiboken version 1.2 have been > > released. PySide is a community supported LGPL wrapper for the Qt > > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > > 1.2 contains numerous stability fixes, improved Windows installation via > > easy_install, and the shiboken module is now correctly installed. > > > > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It > > can be installed via easy_install, pip and windows installer. > > The pypi link points to 1.1.2 and that is all that easy_install finds. > pypi link will be updates soon, thanks > > Joel > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Tue Jul 9 08:35:23 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 08:35:23 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DBA777.5030601@gamry.com> References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> Message-ID: 2013/7/9 Joel B. Mohler > On 7/8/2013 9:30 PM, John Ehresman wrote: > > I'm happy to announce that PySide and Shiboken version 1.2 have been > > released. PySide is a community supported LGPL wrapper for the Qt > > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > > 1.2 contains numerous stability fixes, improved Windows installation via > > easy_install, and the shiboken module is now correctly installed. > > > > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It > > can be installed via easy_install, pip and windows installer. > > The pypi link points to 1.1.2 and that is all that easy_install finds. > PyPI is now updated https://pypi.python.org/pypi/PySide Roman > > Joel > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Tue Jul 9 08:36:51 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 08:36:51 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: References: <51DB67AB.60204@wingware.com> Message-ID: 2013/7/9 Tony Barbieri > Excellent news! > > Are there release notes up anywhere? > The changes are here: https://pypi.python.org/pypi/PySide#changes Regards Roman > > Thanks again for all the hard work! > > -tony > > > On Mon, Jul 8, 2013 at 9:30 PM, John Ehresman wrote: > >> I'm happy to announce that PySide and Shiboken version 1.2 have been >> released. PySide is a community supported LGPL wrapper for the Qt >> libraries and Shiboken is a C++ wrapping tool used by PySide. Version >> 1.2 contains numerous stability fixes, improved Windows installation via >> easy_install, and the shiboken module is now correctly installed. >> >> The pypi page for PySide is at http://pypi.python.org/pypi/PySide It >> can be installed via easy_install, pip and windows installer. >> >> The source tarballs can be found at: >> PySide: >> >> http://download.qt-project.org/official_releases/pyside/pyside-qt4.8+1.2.0.tar.bz2 >> Shiboken: >> >> http://download.qt-project.org/official_releases/pyside/shiboken-1.2.0.tar.bz2 >> >> Questions / feedback may be sent to the pyside mailing list at >> pyside at qt-project.org >> >> Thanks: >> >> Paulo Alcantara >> John Cummings >> Robin Dunn >> John Ehresman >> Roman Lacko >> Stefan Landvogt >> Hugo Parente Lima >> Sébastien Sablé >> Nathan Smith >> Matthew Woehlke >> >> -- >> John Ehresman >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> > > > > -- > -tony > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Tue Jul 9 11:05:32 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Tue, 09 Jul 2013 05:05:32 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> Message-ID: <51DBD25C.5010909@gamry.com> On 7/9/2013 2:35 AM, Roman Lacko wrote: > 2013/7/9 Joel B. Mohler > > > On 7/8/2013 9:30 PM, John Ehresman wrote: > > I'm happy to announce that PySide and Shiboken version 1.2 have been > > released. PySide is a community supported LGPL wrapper for the Qt > > libraries and Shiboken is a C++ wrapping tool used by PySide. > Version > > 1.2 contains numerous stability fixes, improved Windows > installation via > > easy_install, and the shiboken module is now correctly installed. > > > > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide > It > > can be installed via easy_install, pip and windows installer. > > The pypi link points to 1.1.2 and that is all that easy_install finds. > > > PyPI is now updated https://pypi.python.org/pypi/PySide Thank-you Roman. This works with one caveat -- it requires python 2.7.5 and does not work on python 2.7.3. The shiboken-python2.7.dll depends on a symbol '_PyTrash_thread_deposit_object' which evidently doesn't exist in 2.7.3. I personally don't mind upgrading to 2.7.5 and really I'm glad I'm now upgraded .... but it's a bit of a gotcha. This is looking like a very nice PySide release. Thank-you! Joel -------------- next part -------------- An HTML attachment was scrubbed... URL: From jens.lindgren.vfx at gmail.com Tue Jul 9 11:43:00 2013 From: jens.lindgren.vfx at gmail.com (Jens Lindgren) Date: Tue, 9 Jul 2013 11:43:00 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DBD25C.5010909@gamry.com> References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> Message-ID: Great release guys! Would it be possible to add a Python 2.7 64-bit package to the download page? /Jens On Tue, Jul 9, 2013 at 11:05 AM, Joel B. Mohler wrote: > On 7/9/2013 2:35 AM, Roman Lacko wrote: > > 2013/7/9 Joel B. Mohler > >> On 7/8/2013 9:30 PM, John Ehresman wrote: >> > I'm happy to announce that PySide and Shiboken version 1.2 have been >> > released. PySide is a community supported LGPL wrapper for the Qt >> > libraries and Shiboken is a C++ wrapping tool used by PySide. Version >> > 1.2 contains numerous stability fixes, improved Windows installation via >> > easy_install, and the shiboken module is now correctly installed. >> > >> > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It >> > can be installed via easy_install, pip and windows installer. >> >> The pypi link points to 1.1.2 and that is all that easy_install finds. >> > > PyPI is now updated https://pypi.python.org/pypi/PySide > > > Thank-you Roman. This works with one caveat -- it requires python 2.7.5 > and does not work on python 2.7.3. The shiboken-python2.7.dll depends on > a symbol '_PyTrash_thread_deposit_object' which evidently doesn't exist in > 2.7.3. I personally don't mind upgrading to 2.7.5 and really I'm glad I'm > now upgraded .... but it's a bit of a gotcha. > > This is looking like a very nice PySide release. Thank-you! > > Joel > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -- Jens Lindgren -------------------------- Lead Technical Director Magoo 3D Studios -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Tue Jul 9 12:10:58 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 12:10:58 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> Message-ID: 2013/7/9 Jens Lindgren > Great release guys! > Would it be possible to add a Python 2.7 64-bit package to the download > page ? > Yes, the package is prepared and it will be available soon on download page Roman > > /Jens > > > On Tue, Jul 9, 2013 at 11:05 AM, Joel B. Mohler wrote: > >> On 7/9/2013 2:35 AM, Roman Lacko wrote: >> >> 2013/7/9 Joel B. Mohler >> >>> On 7/8/2013 9:30 PM, John Ehresman wrote: >>> > I'm happy to announce that PySide and Shiboken version 1.2 have been >>> > released. PySide is a community supported LGPL wrapper for the Qt >>> > libraries and Shiboken is a C++ wrapping tool used by PySide. Version >>> > 1.2 contains numerous stability fixes, improved Windows installation >>> via >>> > easy_install, and the shiboken module is now correctly installed. >>> > >>> > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It >>> > can be installed via easy_install, pip and windows installer. >>> >>> The pypi link points to 1.1.2 and that is all that easy_install finds. >>> >> >> PyPI is now updated https://pypi.python.org/pypi/PySide >> >> >> Thank-you Roman. This works with one caveat -- it requires python 2.7.5 >> and does not work on python 2.7.3. The shiboken-python2.7.dll depends on >> a symbol '_PyTrash_thread_deposit_object' which evidently doesn't exist in >> 2.7.3. I personally don't mind upgrading to 2.7.5 and really I'm glad I'm >> now upgraded .... but it's a bit of a gotcha. >> >> This is looking like a very nice PySide release. Thank-you! >> >> Joel >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> >> > > > -- > Jens Lindgren > -------------------------- > Lead Technical Director > Magoo 3D Studios > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Tue Jul 9 12:35:50 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Tue, 09 Jul 2013 06:35:50 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 Message-ID: <51DBE786.5080302@gamry.com> Hi, The new path discovery in __init__.py and _utils.py is giving me fits in a frozen environment. I think I have two unrelated problems: 1) It's a known issue that __file__ is not what one typically wants after py2exe or cxfreeze is applied. http://www.py2exe.org/index.cgi/WhereAmI has the canonical alternative (I guess), but I'm not sure if one would want to make the sys.executable assumption in PySide code (???). So I tried modifying my _utils.py just applying the sys.executable alternative and then ... 2) I get a windows segfault on the 'from . import QtCore' line (unfrozen runs fine): Problem Event Name: APPCRASH Application Name: pyside.exe Application Version: 0.0.0.0 Application Timestamp: 50afcb50 Fault Module Name: python27.dll Fault Module Version: 2.7.5150.1013 Fault Module Timestamp: 5193f378 Exception Code: c0000005 Exception Offset: 000ac705 I'm running win7 32 bit. I suppose more could be learned by compiling the debug version, but I'm going to let this stew a bit before I start on that. Thanks for any ideas on approaching this. Joel From jens.lindgren.vfx at gmail.com Tue Jul 9 12:39:57 2013 From: jens.lindgren.vfx at gmail.com (Jens Lindgren) Date: Tue, 9 Jul 2013 12:39:57 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> Message-ID: Awesome!! On Tue, Jul 9, 2013 at 12:10 PM, Roman Lacko wrote: > > 2013/7/9 Jens Lindgren > >> Great release guys! >> Would it be possible to add a Python 2.7 64-bit package to the download >> page ? >> > > Yes, the package is prepared and it will be available soon on download page > > Roman > > >> >> /Jens >> >> >> On Tue, Jul 9, 2013 at 11:05 AM, Joel B. Mohler wrote: >> >>> On 7/9/2013 2:35 AM, Roman Lacko wrote: >>> >>> 2013/7/9 Joel B. Mohler >>> >>>> On 7/8/2013 9:30 PM, John Ehresman wrote: >>>> > I'm happy to announce that PySide and Shiboken version 1.2 have been >>>> > released. PySide is a community supported LGPL wrapper for the Qt >>>> > libraries and Shiboken is a C++ wrapping tool used by PySide. Version >>>> > 1.2 contains numerous stability fixes, improved Windows installation >>>> via >>>> > easy_install, and the shiboken module is now correctly installed. >>>> > >>>> > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide It >>>> > can be installed via easy_install, pip and windows installer. >>>> >>>> The pypi link points to 1.1.2 and that is all that easy_install finds. >>>> >>> >>> PyPI is now updated https://pypi.python.org/pypi/PySide >>> >>> >>> Thank-you Roman. This works with one caveat -- it requires python 2.7.5 >>> and does not work on python 2.7.3. The shiboken-python2.7.dll depends on >>> a symbol '_PyTrash_thread_deposit_object' which evidently doesn't exist in >>> 2.7.3. I personally don't mind upgrading to 2.7.5 and really I'm glad I'm >>> now upgraded .... but it's a bit of a gotcha. >>> >>> This is looking like a very nice PySide release. Thank-you! >>> >>> Joel >>> >>> _______________________________________________ >>> PySide mailing list >>> PySide at qt-project.org >>> http://lists.qt-project.org/mailman/listinfo/pyside >>> >>> >> >> >> -- >> Jens Lindgren >> -------------------------- >> Lead Technical Director >> Magoo 3D Studios >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> >> > -- Jens Lindgren -------------------------- Lead Technical Director Magoo 3D Studios -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Tue Jul 9 13:20:37 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 13:20:37 +0200 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> Message-ID: The PySide Windows build for Python 2.7 64bit is now available on downloads page http://qt-project.org/wiki/PySide_Binaries_Windows Roman 2013/7/9 Jens Lindgren > Awesome!! > > > On Tue, Jul 9, 2013 at 12:10 PM, Roman Lacko wrote: > >> >> 2013/7/9 Jens Lindgren >> >>> Great release guys! >>> Would it be possible to add a Python 2.7 64-bit package to the download >>> page ? >>> >> >> Yes, the package is prepared and it will be available soon on download >> page >> >> Roman >> >> >>> >>> /Jens >>> >>> >>> On Tue, Jul 9, 2013 at 11:05 AM, Joel B. Mohler wrote: >>> >>>> On 7/9/2013 2:35 AM, Roman Lacko wrote: >>>> >>>> 2013/7/9 Joel B. Mohler >>>> >>>>> On 7/8/2013 9:30 PM, John Ehresman wrote: >>>>> > I'm happy to announce that PySide and Shiboken version 1.2 have been >>>>> > released. PySide is a community supported LGPL wrapper for the Qt >>>>> > libraries and Shiboken is a C++ wrapping tool used by PySide. >>>>> Version >>>>> > 1.2 contains numerous stability fixes, improved Windows installation >>>>> via >>>>> > easy_install, and the shiboken module is now correctly installed. >>>>> > >>>>> > The pypi page for PySide is athttp://pypi.python.org/pypi/PySide >>>>> It >>>>> > can be installed via easy_install, pip and windows installer. >>>>> >>>>> The pypi link points to 1.1.2 and that is all that easy_install finds. >>>>> >>>> >>>> PyPI is now updated https://pypi.python.org/pypi/PySide >>>> >>>> >>>> Thank-you Roman. This works with one caveat -- it requires python >>>> 2.7.5 and does not work on python 2.7.3. The shiboken-python2.7.dll >>>> depends on a symbol '_PyTrash_thread_deposit_object' which evidently >>>> doesn't exist in 2.7.3. I personally don't mind upgrading to 2.7.5 and >>>> really I'm glad I'm now upgraded .... but it's a bit of a gotcha. >>>> >>>> This is looking like a very nice PySide release. Thank-you! >>>> >>>> Joel >>>> >>>> _______________________________________________ >>>> PySide mailing list >>>> PySide at qt-project.org >>>> http://lists.qt-project.org/mailman/listinfo/pyside >>>> >>>> >>> >>> >>> -- >>> Jens Lindgren >>> -------------------------- >>> Lead Technical Director >>> Magoo 3D Studios >>> >>> _______________________________________________ >>> PySide mailing list >>> PySide at qt-project.org >>> http://lists.qt-project.org/mailman/listinfo/pyside >>> >>> >> > > > -- > Jens Lindgren > -------------------------- > Lead Technical Director > Magoo 3D Studios > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Tue Jul 9 16:16:11 2013 From: jpe at wingware.com (John Ehresman) Date: Tue, 09 Jul 2013 10:16:11 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DBD25C.5010909@gamry.com> References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> Message-ID: <51DC1B2B.40507@wingware.com> On 7/9/13 5:05 AM, Joel B. Mohler wrote: > Thank-you Roman. This works with one caveat -- it requires python 2.7.5 > and does not work on python 2.7.3. The shiboken-python2.7.dll depends > on a symbol '_PyTrash_thread_deposit_object' which evidently doesn't > exist in 2.7.3. I personally don't mind upgrading to 2.7.5 and really > I'm glad I'm now upgraded .... but it's a bit of a gotcha. This sounds like an unexpected binary compatibility problem (Python doesn't usually add C api's in point releases like this). Maybe we need to be building the binaries with 2.7.0? Thanks for reporting this, John From jpe at wingware.com Tue Jul 9 16:24:18 2013 From: jpe at wingware.com (John Ehresman) Date: Tue, 09 Jul 2013 10:24:18 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DBE786.5080302@gamry.com> References: <51DBE786.5080302@gamry.com> Message-ID: <51DC1D12.30804@wingware.com> On 7/9/13 6:35 AM, Joel B. Mohler wrote: > The new path discovery in __init__.py and _utils.py is giving me fits in > a frozen environment. I think I have two unrelated problems: Would it help to simply disable the path discovery when __file__ doesn't refer to an actual file? I think things should work if all the .pyd's and .dll's were in the right place. Thanks, John From jmohler at gamry.com Tue Jul 9 17:06:03 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Tue, 09 Jul 2013 11:06:03 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DC1D12.30804@wingware.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> Message-ID: <51DC26DB.7080307@gamry.com> On 7/9/2013 10:24 AM, John Ehresman wrote: > On 7/9/13 6:35 AM, Joel B. Mohler wrote: >> The new path discovery in __init__.py and _utils.py is giving me fits in >> a frozen environment. I think I have two unrelated problems: > > Would it help to simply disable the path discovery when __file__ > doesn't refer to an actual file? I think things should work if all > the .pyd's and .dll's were in the right place. I have tried commenting out the _setupQtDirectories in __init__.py. It then crashes (ala segfault) on the import of QtCore. I can't think of anything that makes sense explaining that crash since it works fine unfrozen and the normal cxfreeze failure mode is to have some missing dll or module ... not crashes. Joel From backup.rlacko at gmail.com Tue Jul 9 17:11:27 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Tue, 9 Jul 2013 17:11:27 +0200 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DC26DB.7080307@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> Message-ID: 2013/7/9 Joel B. Mohler > On 7/9/2013 10:24 AM, John Ehresman wrote: > > On 7/9/13 6:35 AM, Joel B. Mohler wrote: > >> The new path discovery in __init__.py and _utils.py is giving me fits in > >> a frozen environment. I think I have two unrelated problems: > > > > Would it help to simply disable the path discovery when __file__ > > doesn't refer to an actual file? I think things should work if all > > the .pyd's and .dll's were in the right place. > > I have tried commenting out the _setupQtDirectories in __init__.py. It > then crashes (ala segfault) on the import of QtCore. I can't think of > anything that makes sense explaining that crash since it works fine > unfrozen and the normal cxfreeze failure mode is to have some missing > dll or module ... not crashes. > When _setupQtDirectories is commented out, i would try to create qt.conf with correct paths. qt.conf should be placed in executable path > > Joel > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Tue Jul 9 17:20:34 2013 From: jpe at wingware.com (John Ehresman) Date: Tue, 09 Jul 2013 11:20:34 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DC26DB.7080307@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> Message-ID: <51DC2A42.6070101@wingware.com> On 7/9/13 11:06 AM, Joel B. Mohler wrote: > On 7/9/2013 10:24 AM, John Ehresman wrote: >> On 7/9/13 6:35 AM, Joel B. Mohler wrote: >>> The new path discovery in __init__.py and _utils.py is giving me fits in >>> a frozen environment. I think I have two unrelated problems: >> >> Would it help to simply disable the path discovery when __file__ >> doesn't refer to an actual file? I think things should work if all >> the .pyd's and .dll's were in the right place. > > I have tried commenting out the _setupQtDirectories in __init__.py. It > then crashes (ala segfault) on the import of QtCore. I can't think of > anything that makes sense explaining that crash since it works fine > unfrozen and the normal cxfreeze failure mode is to have some missing > dll or module ... not crashes. It would be nice for someone to try this in a debug build and hopefull identify where it's segfaulting. Even if it won't work without a qt.conf, it shouldn't segfault. Thanks, John From rdunn at enthought.com Tue Jul 9 23:50:53 2013 From: rdunn at enthought.com (Robin Dunn) Date: Tue, 09 Jul 2013 14:50:53 -0700 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DC1B2B.40507@wingware.com> References: <51DB67AB.60204@wingware.com> <51DBA777.5030601@gamry.com> <51DBD25C.5010909@gamry.com> <51DC1B2B.40507@wingware.com> Message-ID: <51DC85BD.9020904@enthought.com> John Ehresman wrote: > On 7/9/13 5:05 AM, Joel B. Mohler wrote: >> Thank-you Roman. This works with one caveat -- it requires python 2.7.5 >> and does not work on python 2.7.3. The shiboken-python2.7.dll depends >> on a symbol '_PyTrash_thread_deposit_object' which evidently doesn't >> exist in 2.7.3. I personally don't mind upgrading to 2.7.5 and really >> I'm glad I'm now upgraded .... but it's a bit of a gotcha. > > This sounds like an unexpected binary compatibility problem (Python > doesn't usually add C api's in point releases like this). Maybe we need > to be building the binaries with 2.7.0? Or at least 2.7.3, since it has been around for a long time and in my experience is the version that most people still using 2.7 have. -- Robin Dunn Software Engineer Enthought, Inc. rdunn at enthought.com -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Wed Jul 10 18:36:06 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Wed, 10 Jul 2013 12:36:06 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DC2A42.6070101@wingware.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> Message-ID: <51DD8D76.6070502@gamry.com> On 7/9/2013 11:20 AM, John Ehresman wrote: > On 7/9/13 11:06 AM, Joel B. Mohler wrote: >> I have tried commenting out the _setupQtDirectories in __init__.py. It >> then crashes (ala segfault) on the import of QtCore. I can't think of >> anything that makes sense explaining that crash since it works fine >> unfrozen and the normal cxfreeze failure mode is to have some missing >> dll or module ... not crashes. > > It would be nice for someone to try this in a debug build and hopefull > identify where it's segfaulting. Even if it won't work without a > qt.conf, it shouldn't segfault. Aha, this has been an interesting journey to get this all compiled in debug. I have found the problem. Shiboken::AutoDecRef atexit(Shiboken::Module::import("atexit")); Shiboken::AutoDecRef regFunc(PyObject_GetAttrString(atexit, "register")); This C-level module import in qtcore_module_wrapper.cpp imports python module atexit, but it fails and so the second line crashes because atexit is null. So, easy fix is to include "--include-modules=atexit" in the cxfreeze command line. So, what is the correct fix? I think that import in sbkmodule.cpp does right to set the PyErr, but the calling code copied above originating from typesystem_core_common.xml (I think?) doesn't check PyErr_Occurred and I don't know how the error should propagate in the abstract conditions of typesystem_core_common.xml. With that knowledge, then I suggest the following is a cxfreeze friendly implementation of __file__ access appropriate in this context: diff --git a/PySide/_utils.py.in b/PySide/_utils.py.in index fb2d25e..f82ce65 100644 --- a/PySide/_utils.py.in +++ b/PySide/_utils.py.in @@ -84,8 +84,10 @@ if sys.platform == 'win32': return path def get_pyside_dir(): - return _get_win32_case_sensitive_name(os.path.abspath(os.path.dirname(__file__))) + from . import QtCore + return _get_win32_case_sensitive_name(os.path.abspath(os.path.dirname(QtCore.__file__))) else: def get_pyside_dir(): - return os.path.abspath(os.path.dirname(__file__)) + from . import QtCore + return os.path.abspath(os.path.dirname(QtCore.__file__)) I would be happy to hear opinions on these topics. Joel From jmohler at gamry.com Wed Jul 10 18:38:31 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Wed, 10 Jul 2013 12:38:31 -0400 Subject: [PySide] _utils license text Message-ID: <51DD8E07.4040102@gamry.com> Could someone please clarify the license boilerplate in PySide\_utils.py.in? # This file is part of PySide: Python for Qt # # Copyright (C) 2013 Digia Plc and/or its subsidiary(-ies). # # Contact: PySide team # # This program is free software; you can redistribute it and/or # modify it under the terms of the GNU General Public License # version 2 as published by the Free Software Foundation. I thought a major point of PySide is that it was LGPL and not just GPL! Joel From backup.rlacko at gmail.com Wed Jul 10 20:31:54 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Wed, 10 Jul 2013 20:31:54 +0200 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DD8D76.6070502@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> Message-ID: 2013/7/10 Joel B. Mohler > On 7/9/2013 11:20 AM, John Ehresman wrote: > > On 7/9/13 11:06 AM, Joel B. Mohler wrote: > >> I have tried commenting out the _setupQtDirectories in __init__.py. It > >> then crashes (ala segfault) on the import of QtCore. I can't think of > >> anything that makes sense explaining that crash since it works fine > >> unfrozen and the normal cxfreeze failure mode is to have some missing > >> dll or module ... not crashes. > > > > It would be nice for someone to try this in a debug build and hopefull > > identify where it's segfaulting. Even if it won't work without a > > qt.conf, it shouldn't segfault. > > Aha, this has been an interesting journey to get this all compiled in > debug. I have found the problem. > > Shiboken::AutoDecRef atexit(Shiboken::Module::import("atexit")); > Shiboken::AutoDecRef regFunc(PyObject_GetAttrString(atexit, > "register")); > > This C-level module import in qtcore_module_wrapper.cpp imports python > module atexit, but it fails and so the second line crashes because > atexit is null. So, easy fix is to include "--include-modules=atexit" > in the cxfreeze command line. > > So, what is the correct fix? I think that import in sbkmodule.cpp does > right to set the PyErr, but the calling code copied above originating > from typesystem_core_common.xml (I think?) doesn't check PyErr_Occurred > and I don't know how the error should propagate in the abstract > conditions of typesystem_core_common.xml. > > With that knowledge, then I suggest the following is a cxfreeze friendly > implementation of __file__ access appropriate in this context: > > diff --git a/PySide/_utils.py.in b/PySide/_utils.py.in > index fb2d25e..f82ce65 100644 > --- a/PySide/_utils.py.in > +++ b/PySide/_utils.py.in > @@ -84,8 +84,10 @@ if sys.platform == 'win32': > return path > > def get_pyside_dir(): > - return > _get_win32_case_sensitive_name(os.path.abspath(os.path.dirname(__file__))) > + from . import QtCore > + return > > _get_win32_case_sensitive_name(os.path.abspath(os.path.dirname(QtCore.__file__))) > > else: > def get_pyside_dir(): > - return os.path.abspath(os.path.dirname(__file__)) > + from . import QtCore > + return os.path.abspath(os.path.dirname(QtCore.__file__)) > Just small correction - on linux the "import . from QtCore" fails when running post install script first time, so it should be: .... def get_pyside_dir(): from . import QtCore return _get_win32_case_sensitive_name(os.path.abspath(os.path.dirname(QtCore.__file__))) else: def get_pyside_dir(): try: from . import QtCore except ImportError: return os.path.abspath(os.path.dirname(__file__)) else: return os.path.abspath(os.path.dirname(QtCore.__file__)) .... -Roman > I would be happy to hear opinions on these topics. > > Joel > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Wed Jul 10 21:28:57 2013 From: jpe at wingware.com (John Ehresman) Date: Wed, 10 Jul 2013 15:28:57 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DD8D76.6070502@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> Message-ID: <51DDB5F9.3080501@wingware.com> On 7/10/13 12:36 PM, Joel B. Mohler wrote: > Aha, this has been an interesting journey to get this all compiled in > debug. I have found the problem. > > Shiboken::AutoDecRef atexit(Shiboken::Module::import("atexit")); > Shiboken::AutoDecRef regFunc(PyObject_GetAttrString(atexit, > "register")); > > This C-level module import in qtcore_module_wrapper.cpp imports python > module atexit, but it fails and so the second line crashes because > atexit is null. So, easy fix is to include "--include-modules=atexit" > in the cxfreeze command line. Yes, that's the way to avoid triggering the bug. > So, what is the correct fix? I think that import in sbkmodule.cpp does > right to set the PyErr, but the calling code copied above originating > from typesystem_core_common.xml (I think?) doesn't check PyErr_Occurred > and I don't know how the error should propagate in the abstract > conditions of typesystem_core_common.xml. The code needs to check if atexit is NULL and possibly just print a warning and continue without registering the handler. The problem is that segfaults can occur at shutdown (I think) if the handler isn't run. Thanks for tracking this down! John From jpe at wingware.com Wed Jul 10 21:30:16 2013 From: jpe at wingware.com (John Ehresman) Date: Wed, 10 Jul 2013 15:30:16 -0400 Subject: [PySide] _utils license text In-Reply-To: <51DD8E07.4040102@gamry.com> References: <51DD8E07.4040102@gamry.com> Message-ID: <51DDB648.3020705@wingware.com> On 7/10/13 12:38 PM, Joel B. Mohler wrote: > Could someone please clarify the license boilerplate in PySide\_utils.py.in? > > # This file is part of PySide: Python for Qt > # > # Copyright (C) 2013 Digia Plc and/or its subsidiary(-ies). > # > # Contact: PySide team > # > # This program is free software; you can redistribute it and/or > # modify it under the terms of the GNU General Public License > # version 2 as published by the Free Software Foundation. > > I thought a major point of PySide is that it was LGPL and not just GPL! This is a mistake and will be fixed quickly. Thanks, John From jmohler at gamry.com Thu Jul 11 14:43:36 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Thu, 11 Jul 2013 08:43:36 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DDB5F9.3080501@wingware.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> Message-ID: <51DEA878.5010603@gamry.com> On 7/10/2013 3:28 PM, John Ehresman wrote: >> So, what is the correct fix? I think that import in sbkmodule.cpp does >> right to set the PyErr, but the calling code copied above originating >> from typesystem_core_common.xml (I think?) doesn't check PyErr_Occurred >> and I don't know how the error should propagate in the abstract >> conditions of typesystem_core_common.xml. > > The code needs to check if atexit is NULL and possibly just print a > warning and continue without registering the handler. The problem is > that segfaults can occur at shutdown (I think) if the handler isn't run. I'm working up a patch for this. Two questions: What is the correct submission routine? I signed up on gerrit at qt-project.org; do I now just push to git://gitorious.org/pyside/pyside.git and it magically appears for review? How does one do incremental builds of pyside? So far I haven't gotten deeper than "python setup.py build", but that takes ages. Joel From jpe at wingware.com Thu Jul 11 16:36:27 2013 From: jpe at wingware.com (John Ehresman) Date: Thu, 11 Jul 2013 10:36:27 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DEA878.5010603@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> Message-ID: <51DEC2EB.8030306@wingware.com> On 7/11/13 8:43 AM, Joel B. Mohler wrote: >> The code needs to check if atexit is NULL and possibly just print a >> warning and continue without registering the handler. The problem is >> that segfaults can occur at shutdown (I think) if the handler isn't run. > > I'm working up a patch for this. Thanks! > What is the correct submission routine? I signed up on gerrit at > qt-project.org; do I now just push to > git://gitorious.org/pyside/pyside.git and it magically appears for review? I've updated this page on the wiki: http://qt-project.org/wiki/PySide_Development_Getting_Started You may also want to refer to the Qt docs on gerrit: http://qt-project.org/wiki/Gerrit-Introduction http://qt-project.org/wiki/Setting-up-Gerrit > How does one do incremental builds of pyside? So far I haven't gotten > deeper than "python setup.py build", but that takes ages. Roman can probably answer this for the setup build, but in general, updating the typesystem file requires the whole module to be rebuilt. John From backup.rlacko at gmail.com Thu Jul 11 16:39:47 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Thu, 11 Jul 2013 16:39:47 +0200 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DEC2EB.8030306@wingware.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> <51DEC2EB.8030306@wingware.com> Message-ID: 2013/7/11 John Ehresman > On 7/11/13 8:43 AM, Joel B. Mohler wrote: > >> The code needs to check if atexit is NULL and possibly just print a > >> warning and continue without registering the handler. The problem is > >> that segfaults can occur at shutdown (I think) if the handler isn't run. > > > > I'm working up a patch for this. > > Thanks! > > > What is the correct submission routine? I signed up on gerrit at > > qt-project.org; do I now just push to > > git://gitorious.org/pyside/pyside.git and it magically appears for > review? > > I've updated this page on the wiki: > http://qt-project.org/wiki/PySide_Development_Getting_Started > > You may also want to refer to the Qt docs on gerrit: > http://qt-project.org/wiki/Gerrit-Introduction > http://qt-project.org/wiki/Setting-up-Gerrit > > > How does one do incremental builds of pyside? So far I haven't gotten > > deeper than "python setup.py build", but that takes ages. > > Roman can probably answer this for the setup build, but in general, > updating the typesystem file requires the whole module to be rebuilt. > the pyside-setup script does not support incremental builds (yet), you need to manually build pyside and shiboken module if you need incremental builds. -Roman > > John > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From fabcastan at gmail.com Thu Jul 11 17:40:30 2013 From: fabcastan at gmail.com (Fabien Castan) Date: Thu, 11 Jul 2013 17:40:30 +0200 Subject: [PySide] Qt5 support Message-ID: Hi, * * Has someone started some development or test for Qt5 support? * * It could be really usefull to have a basic setup to start the development of Qt5 support. Maybe a project with just the support of QtCore could be a really good starting point. With such setup, external people like me could test and try to help in an iterative process. * * I don’t know what is the difficulty to switch to Qt5. There is always a discussion about improving the binding tool with the binding of Qt5. Are these improvements really needed for Qt5 support? * * The only major change, I have seen, is how signals are managed. There is specific things in PySide and that should be completely invalid in Qt5. But maybe, it’s easier in Qt5 as it uses objects for signals? The other changes are simply a lot of renaming in QtQuick objects, new split between modules and a lot of evolutions in the api. But all that changes are handled in typesystem files and everyone could easily contribute on that part, no? * * I have the habit to use PySide/QML for UI stuff and C++ for the engine. It’s a real pleasure and a big speedup for the development. Using python for the UI is really usefull! But it’s difficult to ignore all the improvements done in Qt5 and QtQuick2. * * Best regards, Fabien -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Thu Jul 11 17:47:40 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Thu, 11 Jul 2013 11:47:40 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DEC2EB.8030306@wingware.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> <51DEC2EB.8030306@wingware.com> Message-ID: <51DED39C.8000004@gamry.com> On 7/11/2013 10:36 AM, John Ehresman wrote: > On 7/11/13 8:43 AM, Joel B. Mohler wrote: >>> The code needs to check if atexit is NULL and possibly just print a >>> warning and continue without registering the handler. The problem is >>> that segfaults can occur at shutdown (I think) if the handler isn't >>> run. >> >> I'm working up a patch for this. > > Thanks! Big thanks to both John & Roman for walking me through this. The patch is up for review, I ran it through on my windows 7 32 bit machine. Roman already gave me the better rendition for Linux in the __file__ changes and the c++ changes look quite platform agnostic to me. Joel From matthew.woehlke at kitware.com Thu Jul 11 18:47:33 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Thu, 11 Jul 2013 12:47:33 -0400 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: <51DEA878.5010603@gamry.com> References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> Message-ID: On 2013-07-11 08:43, Joel B. Mohler wrote: > How does one do incremental builds of pyside? So far I haven't gotten > deeper than "python setup.py build", but that takes ages. I'm using CMake+ninja to build PySide... AFAIK incremental for that should Just Work. (At least it does for other projects I have using shiboken.) You can use any generator you like with CMake; I'm just partial to ninja because it is fast and has good features. As I've used that also to install PySide with success (on Linux at least), I'm not sure why you would need to use the setup.py. -- Matthew From matthew.woehlke at kitware.com Thu Jul 11 19:27:05 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Thu, 11 Jul 2013 13:27:05 -0400 Subject: [PySide] (shiboken) how to wrap template class? Message-ID: I'm trying to write shiboken bindings for some code that makes heavy use of std[::tr1]::shared_ptr, but cannot figure out how to wrap the same. I could've sworn I'd seen shiboken wrapping template stuff somewhere, but can't seem to find any examples (besides the built-in container types, anyway). Can someone recommend a course of action? Thanks, -- Matthew From backup.rlacko at gmail.com Thu Jul 11 19:55:15 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Thu, 11 Jul 2013 19:55:15 +0200 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> Message-ID: 2013/7/11 Matthew Woehlke > On 2013-07-11 08:43, Joel B. Mohler wrote: > > How does one do incremental builds of pyside? So far I haven't gotten > > deeper than "python setup.py build", but that takes ages. > > I'm using CMake+ninja to build PySide... AFAIK incremental for that > should Just Work. (At least it does for other projects I have using > shiboken.) You can use any generator you like with CMake; I'm just > partial to ninja because it is fast and has good features. > > As I've used that also to install PySide with success (on Linux at > least), I'm not sure why you would need to use the setup.py. > Matthew is right about cmake+tools...The pyside-setup project (setup.py) was created for end users who don't need to change pyside and shiboken sources and who wan't just install pyside in system and also for distribution creators (like me). For daily core development it is better to use cmake+whatever tool is neded. -Roman > > -- > Matthew > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Thu Jul 11 20:04:43 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Thu, 11 Jul 2013 20:04:43 +0200 Subject: [PySide] cxfreeze and PySide 1.2.0 In-Reply-To: References: <51DBE786.5080302@gamry.com> <51DC1D12.30804@wingware.com> <51DC26DB.7080307@gamry.com> <51DC2A42.6070101@wingware.com> <51DD8D76.6070502@gamry.com> <51DDB5F9.3080501@wingware.com> <51DEA878.5010603@gamry.com> Message-ID: 2013/7/11 Roman Lacko > 2013/7/11 Matthew Woehlke > >> On 2013-07-11 08:43, Joel B. Mohler wrote: >> > How does one do incremental builds of pyside? So far I haven't gotten >> > deeper than "python setup.py build", but that takes ages. >> >> I'm using CMake+ninja to build PySide... AFAIK incremental for that >> should Just Work. (At least it does for other projects I have using >> shiboken.) You can use any generator you like with CMake; I'm just >> partial to ninja because it is fast and has good features. >> >> As I've used that also to install PySide with success (on Linux at >> least), I'm not sure why you would need to use the setup.py. >> > > Matthew is right about cmake+tools...The pyside-setup project (setup.py) > was created for end users who don't need to change pyside and shiboken > sources and who wan't just install pyside in system and also for > distribution creators (like me). For daily core development it is better to > use cmake+whatever tool is neded. > BTW, regarding to core development, I'm planning to create the buildout [1] enabled project [2] for PySide development. It will help developers to setup complete environmet for Qt and PySide development. [1] https://pypi.python.org/pypi/zc.buildout/2.2.0 [2] https://github.com/PySide/pyside-buildout > -Roman > > >> >> -- >> Matthew >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Thu Jul 11 20:23:05 2013 From: jpe at wingware.com (John Ehresman) Date: Thu, 11 Jul 2013 14:23:05 -0400 Subject: [PySide] (shiboken) how to wrap template class? In-Reply-To: References: Message-ID: <51DEF809.8050403@wingware.com> On 7/11/13 1:27 PM, Matthew Woehlke wrote: > I'm trying to write shiboken bindings for some code that makes heavy use > of std[::tr1]::shared_ptr, but cannot figure out how to wrap the same. > > I could've sworn I'd seen shiboken wrapping template stuff somewhere, > but can't seem to find any examples (besides the built-in container > types, anyway). Can someone recommend a course of action? I don't know if Qt makes use of templates like this; if it does, I would look at the PySide code for examples. I recently had some success using to inject manually written code rather than trying to get arguments to convert, but then I'm more familiar with the Python C api than most and didn't have very many methods to wrap. Cheers, John From jpe at wingware.com Thu Jul 11 21:53:50 2013 From: jpe at wingware.com (John Ehresman) Date: Thu, 11 Jul 2013 15:53:50 -0400 Subject: [PySide] Qt5 support In-Reply-To: References: Message-ID: <51DF0D4E.80307@wingware.com> On 7/11/13 11:40 AM, Fabien Castan wrote: > Has someone started some development or test for Qt5 support? >... > I don’t know what is the difficulty to switch to Qt5. There is always a > discussion about improving the binding tool with the binding of Qt5. Are > these improvements really needed for Qt5 support? We talked about this at the sprint and decided to look at alternative binding tools in conjunction with supporting Qt5. The goal will be a set of bindings that the developers who are currently supporting PySide can more easily support and maintain. None of us were involved with the development of shiboken and we would like to look at alternatives. That said, developer could begin looking at the changes required by Qt 5. PySide is developed and maintained by the community and well thought out patches and other changes to support Qt 5 would likely be accepted. Cheers, John From nmelchior at seegrid.com Thu Jul 11 22:14:21 2013 From: nmelchior at seegrid.com (Nik Melchior) Date: Thu, 11 Jul 2013 16:14:21 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <51DB67AB.60204@wingware.com> References: <51DB67AB.60204@wingware.com> Message-ID: <20130711201420.GA11788@a367.seegrid.com> On Mon, Jul 08, 2013 at 09:30:19PM -0400, John Ehresman wrote: > I'm happy to announce that PySide and Shiboken version 1.2 have been > released. PySide is a community supported LGPL wrapper for the Qt > libraries and Shiboken is a C++ wrapping tool used by PySide. Version > 1.2 contains numerous stability fixes, improved Windows installation via > easy_install, and the shiboken module is now correctly installed. I am happy to see progress, but it's too bad PySide/Tools wasn't released in conjunction with this. There are two open pull requests on github, including my own, which addresses PYSIDE-130 and PYSIDE-131. What's the best step forward? Should the tools be integrated with the Qt repository? I am willing to take a more active role as maintainer of the tools, since my work depends on them. In addition to my github fork, I have a Qt/gerrit account (nmelchior). -- Nik Melchior Sr. Roboticist | 412-379-4500 x147 Seegrid Corporation | www.seegrid.com 216 Park West Drive, Pittsburgh, PA 15275 Email Confidentiality Notice The information contained in this transmission is confidential, proprietary or privileged and may be subject to protection under the law. This message is intended for the sole use of the individual or entity to whom it's addressed. If you are not the intended recipient, you are notified that any use, distribution or copying of the message is strictly prohibited and may subject you to criminal or civil penalties. If you received this transmission in error, please contact the sender immediately by replying to this email and delete the material from any computer. From jpe at wingware.com Thu Jul 11 22:27:19 2013 From: jpe at wingware.com (John Ehresman) Date: Thu, 11 Jul 2013 16:27:19 -0400 Subject: [PySide] Move PySide development to github? Message-ID: <51DF1527.3020003@wingware.com> At the sprint in Austin, a few of us discussed moving the git repository where changes are made to github. We would still limit who could push to a core group of approvers and require someone else to review changes in almost all cases. The goal is to move away from gerrit / gitorious to the github tools. Each of us has limited time to contribute to PySide and we'd rather spend that time improving PySide rather than learning how to use tools we don't otherwise use. What to do with the bug tracker is less clear. It probably would migrate to github's tracker or another tracker because the current tracker doesn't work well, but OTOH no one is particularly impressed with github's tracker. We decided to float this idea on email and see what the response was. Any transition wouldn't happen for a few weeks and all patches should be submitted via Gerrit until then. Cheers, John From matthew.woehlke at kitware.com Thu Jul 11 23:04:09 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Thu, 11 Jul 2013 17:04:09 -0400 Subject: [PySide] (shiboken) how to wrap template class? In-Reply-To: <51DEF809.8050403@wingware.com> References: <51DEF809.8050403@wingware.com> Message-ID: On 2013-07-11 14:23, John Ehresman wrote: > On 7/11/13 1:27 PM, Matthew Woehlke wrote: >> I'm trying to write shiboken bindings for some code that makes heavy use >> of std[::tr1]::shared_ptr, but cannot figure out how to wrap the same. >> >> I could've sworn I'd seen shiboken wrapping template stuff somewhere, >> but can't seem to find any examples (besides the built-in container >> types, anyway). Can someone recommend a course of action? > > I don't know if Qt makes use of templates like this; if it does, I would > look at the PySide code for examples. I couldn't find any; I was hoping someone on the list might be able to point to one :-). > I recently had some success using to inject manually > written code rather than trying to get arguments to convert, but then > I'm more familiar with the Python C api than most and didn't have very > many methods to wrap. So... I tried getting Shiboken to wrap a simple template class, and it ended badly. I suspect Shiboken may not 'really' support template classes at all except for known container types. I suppose if I was feeling ambitious, I could add a pointer container type. That said, what I got to work for me is to typedef all of the shared pointer usages and to wrap the typedefs. This got me wrapped classes but with no methods. However I was able to add: %RETURN_TYPE the_ptr = %CPPSELF->get(); %PYARG_0 = %CONVERTTOPYTHON[%RETURN_TYPE](the_ptr); ...which seems to be sufficient for something usable and to understand generally how to get it working. It's too bad I can't template away adding the method, or even the entire wrapping of the pointer :-). -- Matthew From jpe at wingware.com Fri Jul 12 00:04:39 2013 From: jpe at wingware.com (John Ehresman) Date: Thu, 11 Jul 2013 18:04:39 -0400 Subject: [PySide] PySide and Shiboken 1.2 released In-Reply-To: <20130711201420.GA11788@a367.seegrid.com> References: <51DB67AB.60204@wingware.com> <20130711201420.GA11788@a367.seegrid.com> Message-ID: <51DF2BF7.7060109@wingware.com> On 7/11/13 4:14 PM, Nik Melchior wrote: > am happy to see progress, but it's too bad PySide/Tools wasn't released in > conjunction with this. There are two open pull requests on github, including > my own, which addresses PYSIDE-130 and PYSIDE-131. Sorry for not reviewing the PySide Tools pull requests. I do think all the repositories should be in one place, probably on github. > What's the best step forward? Should the tools be integrated with the Qt > repository? I am willing to take a more active role as maintainer of the > tools, since my work depends on them. In addition to my github fork, I have > a Qt/gerrit account (nmelchior). Thanks for offering to help. I'd like to defer figuring out how to do this until we have all the repositories in one place. John From felix.morency at gmail.com Fri Jul 12 03:59:31 2013 From: felix.morency at gmail.com (=?UTF-8?Q?F=C3=A9lix_C=2E_Morency?=) Date: Thu, 11 Jul 2013 21:59:31 -0400 Subject: [PySide] Move PySide development to github? In-Reply-To: <51DF1527.3020003@wingware.com> References: <51DF1527.3020003@wingware.com> Message-ID: Can you elaborate on the reasons no one is impressed with the Github tracker? On Thu, Jul 11, 2013 at 4:27 PM, John Ehresman wrote: > At the sprint in Austin, a few of us discussed moving the git repository > where changes are made to github. We would still limit who could push > to a core group of approvers and require someone else to review changes > in almost all cases. The goal is to move away from gerrit / gitorious > to the github tools. Each of us has limited time to contribute to > PySide and we'd rather spend that time improving PySide rather than > learning how to use tools we don't otherwise use. > > What to do with the bug tracker is less clear. It probably would > migrate to github's tracker or another tracker because the current > tracker doesn't work well, but OTOH no one is particularly impressed > with github's tracker. > > We decided to float this idea on email and see what the response was. > Any transition wouldn't happen for a few weeks and all patches should be > submitted via Gerrit until then. > > Cheers, > > John > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- Félix C. Morency, M.Sc. Plateforme d’analyse et de visualisation d’images Centre Hospitalier Universitaire de Sherbrooke Centre de recherche clinique Étienne-Le Bel Local Z5-3031 | 819.346.1110 ext 16634 -------------- next part -------------- An HTML attachment was scrubbed... URL: From sean at seanfisk.com Fri Jul 12 05:30:56 2013 From: sean at seanfisk.com (Sean Fisk) Date: Thu, 11 Jul 2013 21:30:56 -0600 Subject: [PySide] Move PySide development to github? In-Reply-To: <51DF1527.3020003@wingware.com> References: <51DF1527.3020003@wingware.com> Message-ID: Moving development to Github sounds like a positive idea to me. A lot of open-source projects are developed there now, and many gain a lot of traction due to the fact they are hosted on the site. I personally think I would be more likely to contribute because I am already used to the Github workflow and know how to use it. Just my 2¢. On Thu, Jul 11, 2013 at 2:27 PM, John Ehresman wrote: > At the sprint in Austin, a few of us discussed moving the git repository > where changes are made to github. We would still limit who could push > to a core group of approvers and require someone else to review changes > in almost all cases. The goal is to move away from gerrit / gitorious > to the github tools. Each of us has limited time to contribute to > PySide and we'd rather spend that time improving PySide rather than > learning how to use tools we don't otherwise use. > > What to do with the bug tracker is less clear. It probably would > migrate to github's tracker or another tracker because the current > tracker doesn't work well, but OTOH no one is particularly impressed > with github's tracker. > > We decided to float this idea on email and see what the response was. > Any transition wouldn't happen for a few weeks and all patches should be > submitted via Gerrit until then. > > Cheers, > > John > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -- Sean Fisk -------------- next part -------------- An HTML attachment was scrubbed... URL: From mairas at iki.fi Fri Jul 12 10:16:00 2013 From: mairas at iki.fi (Matti Airas) Date: Fri, 12 Jul 2013 11:16:00 +0300 Subject: [PySide] Move PySide development to github? In-Reply-To: <51DF1527.3020003@wingware.com> References: <51DF1527.3020003@wingware.com> Message-ID: Hi, Just note that moving from Qt's Gerrit to Github means moving away from the Qt Project: if the contributors don't go through the license agreement click-through at Gerrit, code pushed to Github cannot later on be merged back to Github. Having said that, I agree that Github works well and is used very widely in the community, and since Digia seems to have no interest in supporting PySide, the move probably would have a net positive effect for PySide development. Just wanted to make sure you're making a conscious decision about this. :-) Cheers, ma. On 11 July 2013 23:27, John Ehresman wrote: > At the sprint in Austin, a few of us discussed moving the git repository > where changes are made to github. We would still limit who could push > to a core group of approvers and require someone else to review changes > in almost all cases. The goal is to move away from gerrit / gitorious > to the github tools. Each of us has limited time to contribute to > PySide and we'd rather spend that time improving PySide rather than > learning how to use tools we don't otherwise use. > > What to do with the bug tracker is less clear. It probably would > migrate to github's tracker or another tracker because the current > tracker doesn't work well, but OTOH no one is particularly impressed > with github's tracker. > > We decided to float this idea on email and see what the response was. > Any transition wouldn't happen for a few weeks and all patches should be > submitted via Gerrit until then. > > Cheers, > > John > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From srekel at gmail.com Fri Jul 12 13:42:01 2013 From: srekel at gmail.com (Anders Elfgren) Date: Fri, 12 Jul 2013 13:42:01 +0200 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries Message-ID: Hi there, I want to try out PySide to build some internal apps for our company, but it seems like the only binaries available on Qt are for 4.8.5, and that the currently downloadable version of PySide is built against 4.8.4. I'm not *sure* this is the reason, but I get this and I'm not sure what else it could be: >>> import PySide >>> import PySide.QtCore Traceback (most recent call last): File "", line 1, in ImportError: DLL load failed: The specified procedure could not be found. I would rather not build PySide and Qt myself if possible. In the Qt archives (http://download.qt-project.org/archive/qt/ ) there are downloads for 5.0 and everything from 4.7 and below, but not the 4.8 builds, which is rather unfortunate. Any ideas on how I can best get ahold of those binaries and give PySide a try? :) Thanks, Anders -- ------------------------------------------------------- Anders Elfgren, AKA Srekel -------------- next part -------------- An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Fri Jul 12 13:55:49 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Fri, 12 Jul 2013 13:55:49 +0200 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries In-Reply-To: References: Message-ID: 2013/7/12 Anders Elfgren > Hi there, > > I want to try out PySide to build some internal apps for our company, but > it seems like the only binaries available on Qt are for 4.8.5, and that the > currently downloadable version of PySide is built against 4.8.4. > here is the proper url where you can download all 4.8 binaries download.qt-project.org/official_releases/qt/4.8 > > I'm not *sure* this is the reason, but I get this and I'm not sure what > else it could be: > The Qt 4.8.5 was released few hours before we have created the PySde windows binaries, there was no time to test the PySide against new Qt release > >>> import PySide > >>> import PySide.QtCore > Traceback (most recent call last): > File "", line 1, in > ImportError: DLL load failed: The specified procedure could not be found. > > > I would rather not build PySide and Qt myself if possible. In the Qt > archives (http://download.qt-project.org/archive/qt/ ) there are > downloads for 5.0 and everything from 4.7 and below, but not the 4.8 > builds, which is rather unfortunate. > > Any ideas on how I can best get ahold of those binaries and give PySide a > try? :) > > Thanks, > > Anders > > > -- > ------------------------------------------------------- > Anders Elfgren, AKA Srekel > > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From srekel at gmail.com Fri Jul 12 14:13:08 2013 From: srekel at gmail.com (Anders Elfgren) Date: Fri, 12 Jul 2013 14:13:08 +0200 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries In-Reply-To: References: Message-ID: Ah, thank you! I googled like crazy for those binaries :) Unfortunately, it doesn't seem to have helped - I installed, updated PATH to point to 4.8.4/bin, and still get the same error message. Could it be that I need the exact same binaries - I downloaded the ones built with MSVC2010 ( qt-win-opensource-4.8.4-vs2010.exe)? If so, which ones do you think it would be? Best regards, Anders On Fri, Jul 12, 2013 at 1:55 PM, Roman Lacko wrote: > 2013/7/12 Anders Elfgren > >> Hi there, >> >> I want to try out PySide to build some internal apps for our company, but >> it seems like the only binaries available on Qt are for 4.8.5, and that the >> currently downloadable version of PySide is built against 4.8.4. >> > > here is the proper url where you can download all 4.8 binaries > download.qt-project.org/official_releases/qt/4.8 > > > >> >> I'm not *sure* this is the reason, but I get this and I'm not sure what >> else it could be: >> > > The Qt 4.8.5 was released few hours before we have created the PySde > windows binaries, there was no time to test the PySide against new Qt > release > > >> >>> import PySide >> >>> import PySide.QtCore >> Traceback (most recent call last): >> File "", line 1, in >> ImportError: DLL load failed: The specified procedure could not be found. >> >> >> I would rather not build PySide and Qt myself if possible. In the Qt >> archives (http://download.qt-project.org/archive/qt/ ) there are >> downloads for 5.0 and everything from 4.7 and below, but not the 4.8 >> builds, which is rather unfortunate. >> >> Any ideas on how I can best get ahold of those binaries and give PySide a >> try? :) >> >> Thanks, >> >> Anders >> >> >> -- >> ------------------------------------------------------- >> Anders Elfgren, AKA Srekel >> >> >> _______________________________________________ >> PySide mailing list >> PySide at qt-project.org >> http://lists.qt-project.org/mailman/listinfo/pyside >> >> > -- ------------------------------------------------------- Anders Elfgren, AKA Srekel -------------- next part -------------- An HTML attachment was scrubbed... URL: From jmohler at gamry.com Fri Jul 12 14:35:44 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Fri, 12 Jul 2013 08:35:44 -0400 Subject: [PySide] Move PySide development to github? In-Reply-To: References: <51DF1527.3020003@wingware.com> Message-ID: <51DFF820.6090900@gamry.com> On 7/12/2013 4:16 AM, Matti Airas wrote: > Hi, > > Just note that moving from Qt's Gerrit to Github means moving away > from the Qt Project: if the contributors don't go through the license > agreement click-through at Gerrit, code pushed to Github cannot later > on be merged back to Github. I think this legal copyright question is the more difficult question for the project. On a purely technical basis, I'd rather deal with github by a mile -- having just figured out gerrit and finding it quite annoying. The gerrit code review is maybe nice, but I'd far rather see code review discussions on the mailing list. I can't learn from other folks' code reviews if I never see them. Joel From backup.rlacko at gmail.com Fri Jul 12 15:01:49 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Fri, 12 Jul 2013 15:01:49 +0200 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries In-Reply-To: References: Message-ID: 2013/7/12 Anders Elfgren > Ah, thank you! I googled like crazy for those binaries :) > > Unfortunately, it doesn't seem to have helped - I installed, updated PATH > to point to 4.8.4/bin, and still get the same error message. > pyside windows binaries are self contained, that means you don't have to install Qt or set PATH, PySide allready contains all required Qt dlls. Could it be that I need the exact same binaries - I downloaded the ones > built with MSVC2010 ( qt-win-opensource-4.8.4-vs2010.exe)? > If so, which ones do you think it would be? > Pyside for python 2.7 is build against Qt 4.8.4-msvc-2008 Pyside for python 3.3 is build against Qt 4.8.4-msvc-2010 Best regards, > > Anders > > > On Fri, Jul 12, 2013 at 1:55 PM, Roman Lacko wrote: > >> 2013/7/12 Anders Elfgren >> >>> Hi there, >>> >>> I want to try out PySide to build some internal apps for our company, >>> but it seems like the only binaries available on Qt are for 4.8.5, and that >>> the currently downloadable version of PySide is built against 4.8.4. >>> >> >> here is the proper url where you can download all 4.8 binaries >> download.qt-project.org/official_releases/qt/4.8 >> >> >> >>> >>> I'm not *sure* this is the reason, but I get this and I'm not sure what >>> else it could be: >>> >> >> The Qt 4.8.5 was released few hours before we have created the PySde >> windows binaries, there was no time to test the PySide against new Qt >> release >> >> >>> >>> import PySide >>> >>> import PySide.QtCore >>> Traceback (most recent call last): >>> File "", line 1, in >>> ImportError: DLL load failed: The specified procedure could not be found. >>> >>> >>> I would rather not build PySide and Qt myself if possible. In the Qt >>> archives (http://download.qt-project.org/archive/qt/ ) there are >>> downloads for 5.0 and everything from 4.7 and below, but not the 4.8 >>> builds, which is rather unfortunate. >>> >>> Any ideas on how I can best get ahold of those binaries and give PySide >>> a try? :) >>> >>> Thanks, >>> >>> Anders >>> >>> >>> -- >>> ------------------------------------------------------- >>> Anders Elfgren, AKA Srekel >>> >>> >>> _______________________________________________ >>> PySide mailing list >>> PySide at qt-project.org >>> http://lists.qt-project.org/mailman/listinfo/pyside >>> >>> >> > > > -- > ------------------------------------------------------- > Anders Elfgren, AKA Srekel > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From srekel at gmail.com Fri Jul 12 15:15:13 2013 From: srekel at gmail.com (Anders Elfgren) Date: Fri, 12 Jul 2013 15:15:13 +0200 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries In-Reply-To: References: Message-ID: Oh I see, the documentation is a bit misleading then. I removed Qt from the path now. I wonder what could then be the reason for me getting the problem? I tried reinstalling now (I had missed that I needed to run distribute_setup.py first) and noticed that I got this message at the end of the installation: Traceback (most recent call last): File "", line 242, in install_win32 ImportError: DLL load failed: The specified procedure could not be found. close failed in file object destructor: sys.excepthook is missing lost sys.stderr I tried installing Python 3.3 and the matching PySide version and that seemed to work - no errors when running import PySide.QtCore :) A bit annoying but I can live with using Python3, at least for the moment. :) If there's anything I can do to get more info on what's causing the problem I'd be glad to try it out. /Anders On Fri, Jul 12, 2013 at 3:01 PM, Roman Lacko wrote: > 2013/7/12 Anders Elfgren > >> Ah, thank you! I googled like crazy for those binaries :) >> >> Unfortunately, it doesn't seem to have helped - I installed, updated PATH >> to point to 4.8.4/bin, and still get the same error message. >> > > pyside windows binaries are self contained, that means you don't have to > install Qt or set PATH, PySide allready contains all required Qt dlls. > > > Could it be that I need the exact same binaries - I downloaded the ones >> built with MSVC2010 ( qt-win-opensource-4.8.4-vs2010.exe)? >> If so, which ones do you think it would be? >> > > Pyside for python 2.7 is build against Qt 4.8.4-msvc-2008 > Pyside for python 3.3 is build against Qt 4.8.4-msvc-2010 > > Best regards, >> >> Anders >> >> >> On Fri, Jul 12, 2013 at 1:55 PM, Roman Lacko wrote: >> >>> 2013/7/12 Anders Elfgren >>> >>>> Hi there, >>>> >>>> I want to try out PySide to build some internal apps for our company, >>>> but it seems like the only binaries available on Qt are for 4.8.5, and that >>>> the currently downloadable version of PySide is built against 4.8.4. >>>> >>> >>> here is the proper url where you can download all 4.8 binaries >>> download.qt-project.org/official_releases/qt/4.8 >>> >>> >>> >>>> >>>> I'm not *sure* this is the reason, but I get this and I'm not sure what >>>> else it could be: >>>> >>> >>> The Qt 4.8.5 was released few hours before we have created the PySde >>> windows binaries, there was no time to test the PySide against new Qt >>> release >>> >>> >>>> >>> import PySide >>>> >>> import PySide.QtCore >>>> Traceback (most recent call last): >>>> File "", line 1, in >>>> ImportError: DLL load failed: The specified procedure could not be >>>> found. >>>> >>>> >>>> I would rather not build PySide and Qt myself if possible. In the Qt >>>> archives (http://download.qt-project.org/archive/qt/ ) there are >>>> downloads for 5.0 and everything from 4.7 and below, but not the 4.8 >>>> builds, which is rather unfortunate. >>>> >>>> Any ideas on how I can best get ahold of those binaries and give PySide >>>> a try? :) >>>> >>>> Thanks, >>>> >>>> Anders >>>> >>>> >>>> -- >>>> ------------------------------------------------------- >>>> Anders Elfgren, AKA Srekel >>>> >>>> >>>> _______________________________________________ >>>> PySide mailing list >>>> PySide at qt-project.org >>>> http://lists.qt-project.org/mailman/listinfo/pyside >>>> >>>> >>> >> >> >> -- >> ------------------------------------------------------- >> Anders Elfgren, AKA Srekel >> >> > -- ------------------------------------------------------- Anders Elfgren, AKA Srekel -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Fri Jul 12 17:14:34 2013 From: jpe at wingware.com (John Ehresman) Date: Fri, 12 Jul 2013 11:14:34 -0400 Subject: [PySide] Getting ahold of Qt 4.8.4 Windows binaries In-Reply-To: References: Message-ID: <51E01D5A.1010306@wingware.com> On 7/12/13 9:15 AM, Anders Elfgren wrote: > I tried reinstalling now (I had missed that I needed to run > distribute_setup.py first) and noticed that I got this message at the > end of the installation: > > Traceback (most recent call last): > File "", line 242, in install_win32 > ImportError: DLL load failed: The specified procedure could not be found. > close failed in file object destructor: > sys.excepthook is missing > lost sys.stderr What is the full version of the Python 2.7 you are using? The binaries were built with 2.7.5 (I think) and Python 2.7.3 & earlier can't use them. The plan is to release a new version that is 2.7.3 compatible soon. Thanks, John From jpe at wingware.com Fri Jul 12 17:28:49 2013 From: jpe at wingware.com (John Ehresman) Date: Fri, 12 Jul 2013 11:28:49 -0400 Subject: [PySide] 1.2.1 release early next week Message-ID: <51E020B1.9060807@wingware.com> Just a heads up -- I want to release a 1.2.1 early next week, primarily so the Windows binary installer can be rebuilt to be compatible with Python 2.7.3. The license notice in _utils.py will also be fixed and a few other small fixes may go in. Cheers, John From jpe at wingware.com Fri Jul 12 17:49:29 2013 From: jpe at wingware.com (John Ehresman) Date: Fri, 12 Jul 2013 11:49:29 -0400 Subject: [PySide] Move PySide development to github? In-Reply-To: References: <51DF1527.3020003@wingware.com> Message-ID: <51E02589.5070002@wingware.com> On 7/12/13 4:16 AM, Matti Airas wrote: > Just note that moving from Qt's Gerrit to Github means moving away from > the Qt Project: if the contributors don't go through the license > agreement click-through at Gerrit, code pushed to Github cannot later on > be merged back to Github. Yes, this is correct: to maintain the current code licensing scheme we'd need to make sure anyone who contributes code to Github had clicked through the license agreement on Gerrit. I like to switch to Github even if we do this. I could also see opening thing up a bit more with PySide so that simple fixes could be accepted from the bug tracker (yes, I realize that there's no clear definition of "simple"). > Having said that, I agree that Github works well and is used very widely > in the community, and since Digia seems to have no interest in > supporting PySide, the move probably would have a net positive effect > for PySide development. Just wanted to make sure you're making a > conscious decision about this. :-) While I appreciate the help that the Qt project has given us, I don't see it as a requirement that PySide remain under the Qt project umbrella. Cheers, John From jpe at wingware.com Fri Jul 12 22:01:26 2013 From: jpe at wingware.com (John Ehresman) Date: Fri, 12 Jul 2013 16:01:26 -0400 Subject: [PySide] Move PySide development to github? In-Reply-To: <20130712172522.5128332.71113.1383@digia.com> References: <51DF1527.3020003@wingware.com> <20130712172522.5128332.71113.1383@digia.com> Message-ID: <51E06096.2020804@wingware.com> On 7/12/13 1:25 PM, Knoll Lars wrote: > Matti is correct, that moving to git hub would also imply moving away > from qt-project. Personally i'd see this as a big loss, as I was hoping > to be able to bring the Python bindings to Qt closer to the core > product in the long term. > > You're right in saying that Digia hasn't done a lot to support pyside > (apart from financing qt-project). Given that Digia had to focus on > securing the future of the core product first that shouldn't come as a > surprise. > > There is a possibility that this will change moving forward, but I can't > promise anything right now. Currently most people in Finland and Norway > (including myself) are on vacations, but I'd like to pick this up during > the contributor summit. If nobody from pyside is there in person, how > about a chat on irc or by phone? I won't be at the contributor summit and I don't believe anyone else working on PySide will be either. I'd like to chat by irc or phone (irc is probably preferred so others can join in). We could try to do it during the summit, which runs from next Monday through Wednesday. Thanks, John From pygestionmgd at gmail.com Sun Jul 14 21:32:27 2013 From: pygestionmgd at gmail.com (=?ISO-8859-1?Q?Jes=FAs?=) Date: Sun, 14 Jul 2013 21:32:27 +0200 Subject: [PySide] QSslSocket error Message-ID: <51E2FCCA.3030700@gmail.com> Hi, With the 1.2.0 version installed in windowsxp , when I exec any pyside-example, appears the next message : QSslSocket: cannot resolve OPENSSL_add_all_algorithms_noconf QSslSocket: cannot resolve OPENSSL_add_all_algorithms_conf QSslSocket: cannot call unresolved function OPENSSL_add_all_algorithms_noconf Someone knows what is the problem ? Cheers Jesús From pygestionmgd at gmail.com Mon Jul 15 11:52:48 2013 From: pygestionmgd at gmail.com (Jesus Martinez) Date: Mon, 15 Jul 2013 11:52:48 +0200 Subject: [PySide] QSslSocket error Message-ID: Sorry, problem fixed with a new installation from 0. 2013/7/14 Jesús > Hi, > > With the 1.2.0 version installed in windowsxp , when I exec any > pyside-example, appears the next message : > > > QSslSocket: cannot resolve OPENSSL_add_all_algorithms_**noconf > QSslSocket: cannot resolve OPENSSL_add_all_algorithms_**conf > QSslSocket: cannot call unresolved function OPENSSL_add_all_algorithms_** > noconf > > > Someone knows what is the problem ? > > Cheers > Jesús > -------------- next part -------------- An HTML attachment was scrubbed... URL: From vihorev at gmail.com Mon Jul 15 14:57:17 2013 From: vihorev at gmail.com (Alexey Vihorev) Date: Mon, 15 Jul 2013 15:57:17 +0300 Subject: [PySide] QtCore.QLibraryInfo.location returns a path which does not even exist Message-ID: Hi! I've hit what looks like a bug here: QtCore.QLibraryInfo.location(QtCore.QLibraryInfo.BinariesPath) prints 'C:\\Qt\\qt-4.8.4-msvc2010-x32\\bin', and this machine I'm on does not even have this directory. As a result of this PySide does not see translation folder, plugin folders, etc. Is there a way to work around this? -------------- next part -------------- An HTML attachment was scrubbed... URL: From payaam.shivaa at gmail.com Fri Jul 19 10:35:42 2013 From: payaam.shivaa at gmail.com (Payam Shiva) Date: Fri, 19 Jul 2013 13:05:42 +0430 Subject: [PySide] Installing Qt Quick Controls Message-ID: Hi, I need to use Qt Quick Controls (or Qt Desktop Components) with PySide. I can't find an installation guide for PySide. Could you help me with this? Thanks -------------- next part -------------- An HTML attachment was scrubbed... URL: From khertan at khertan.net Fri Jul 19 13:44:50 2013 From: khertan at khertan.net (khertan at khertan.net) Date: Fri, 19 Jul 2013 11:44:50 +0000 Subject: [PySide] Installing Qt Quick Controls In-Reply-To: References: Message-ID: Hi, that's part of Qt5 if i m right ... PySide isn't Qt5 ready ... -- Benoît HERVIER - http://khertan.netLe 19/07/13 10:35 Payam Shiva a écrit : Hi, I need to use Qt Quick Controls (or Qt Desktop Components) with PySide. I can't find an installation guide for PySide. Could you help me with this? Thanks From gargi.das at hotmail.com Fri Jul 19 14:23:43 2013 From: gargi.das at hotmail.com (Gargi Das) Date: Fri, 19 Jul 2013 17:53:43 +0530 Subject: [PySide] Installing Qt Quick Controls Message-ID: Yes absolutely PySide is still not Qt5 so we all have to wait for that and the qt controls !! Sent from my Windows Phone ________________________________ From: khertan at khertan.net Sent: ‎19-‎07-‎2013 05:15 PM To: pyside at qt-project.org; Payam Shiva Subject: Re: [PySide] Installing Qt Quick Controls Hi, that's part of Qt5 if i m right ... PySide isn't Qt5 ready ... -- Benoît HERVIER - http://khertan.netLe 19/07/13 10:35 Payam Shiva a écrit : Hi, I need to use Qt Quick Controls (or Qt Desktop Components) with PySide. I can't find an installation guide for PySide. Could you help me with this? Thanks _______________________________________________ PySide mailing list PySide at qt-project.org http://lists.qt-project.org/mailman/listinfo/pyside -------------- next part -------------- An HTML attachment was scrubbed... URL: From payaam.shivaa at gmail.com Fri Jul 19 14:36:37 2013 From: payaam.shivaa at gmail.com (Payam Shiva) Date: Fri, 19 Jul 2013 17:06:37 +0430 Subject: [PySide] Installing Qt Quick Controls In-Reply-To: References: Message-ID: The newest version is based on Qt Quick 2.0 thus it only works on Qt5. But Desktop Components have been in development for a long time and earlier versions should work on Qt4 as well. There is a StackOverflow answer[1] explaining how to install it as a Qt 4.7 plugin, but I don't understand how to translate it to PySide. I use the binary package of PySide 1.2 on windows 7. Should I start with Qt 4.8 source, compile it, compile and install desktop components, then compile PySide? Or can I use binary packages for Qt and PySide and just compile desktop components? Which steps should I take to achieve that? In any situation, where can I find the latest version working with Qt4? [1] http://stackoverflow.com/questions/6298584/installing-qt-quick-components-for-desktop-for-use-with-qt-creator On Fri, Jul 19, 2013 at 4:14 PM, wrote: > Hi, that's part of Qt5 if i m right ... > PySide isn't Qt5 ready ... > > -- > Benoît HERVIER - http://khertan.netLe 19/07/13 10:35 Payam Shiva a écrit : > Hi, > > > I need to use Qt Quick Controls (or Qt Desktop Components) with PySide. I > can't find an installation guide for PySide. Could you help me with this? > > > Thanks > -------------- next part -------------- An HTML attachment was scrubbed... URL: From robert.schilling at gmx.at Sun Jul 21 17:56:04 2013 From: robert.schilling at gmx.at (Robert Schilling) Date: Sun, 21 Jul 2013 17:56:04 +0200 (CEST) Subject: [PySide] PySide 1.2 for Windows Message-ID: As far as I can see there is only a Python 2.7 and Python 3.3 releaseo for PySide 1.2. Will there also be a Python 3.2 release as in the past? Thank you in advance! Regards Robert From robert.schilling at gmx.at Sun Jul 21 19:08:31 2013 From: robert.schilling at gmx.at (Robert Schilling) Date: Sun, 21 Jul 2013 19:08:31 +0200 (CEST) Subject: [PySide] Installing PySide 1.2 on Windows via easy_install fails Message-ID: An HTML attachment was scrubbed... URL: From backup.rlacko at gmail.com Sun Jul 21 20:03:58 2013 From: backup.rlacko at gmail.com (Roman Lacko) Date: Sun, 21 Jul 2013 20:03:58 +0200 Subject: [PySide] Installing PySide 1.2 on Windows via easy_install fails In-Reply-To: References: Message-ID: Hi, what python version are you using ? The windows binaries are released only for Python 2.7 and 3.3. Regards Roman 2013/7/21 Robert Schilling > I'm trying to install PySide via easy_install into my virtual environment. > As I thought this installs a binary releas of PyVisa. > Instead I get the following errors, which seems to try to build PySide > locally: > > (venv) C:\data\repos\myproj>easy_install PySide > Searching for PySide > Reading http://pypi.python.org/simple/PySide/ > Reading http://www.pyside.org > Reading http://releases.qt-project.org/pyside/1.1.1/ > Reading http://www.pyside.org/files/ > Reading http://www.pyside.org/files/pkg/ > Reading http://releases.qt-project.org/pyside/ > Reading http://download.qt-project.org/official_releases/pyside/ > Best match: PySide 1.2.0 > Downloading > https://pypi.python.org/packages/source/P/PySide/PySide-1.2.0.tar.gz > #md5=a6f8ce6c0ab4ce858a6bdd10a3af3101 > Processing PySide-1.2.0.tar.gz > Writing > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\setup.cfg > Running PySide-1.2.0\setup.py -q bdist_egg --dist-dir > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\egg-dist-tmp-tj2oiv > Removing > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\pyside_package > error: Setup script exited with error: Failed to find qmake. Please > specify thepath to qmake with --qmake parameter. > > Am I doing something wrong? > > Regarsd Robert > > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From fred.f.bat at gmail.com Mon Jul 22 06:11:10 2013 From: fred.f.bat at gmail.com (frederico Batista) Date: Mon, 22 Jul 2013 01:11:10 -0300 Subject: [PySide] Installing PySide 1.2 on Windows via easy_install fails In-Reply-To: References: Message-ID: Did you install qmake? On Jul 21, 2013 2:08 PM, "Robert Schilling" wrote: > I'm trying to install PySide via easy_install into my virtual environment. > As I thought this installs a binary releas of PyVisa. > Instead I get the following errors, which seems to try to build PySide > locally: > > (venv) C:\data\repos\myproj>easy_install PySide > Searching for PySide > Reading http://pypi.python.org/simple/PySide/ > Reading http://www.pyside.org > Reading http://releases.qt-project.org/pyside/1.1.1/ > Reading http://www.pyside.org/files/ > Reading http://www.pyside.org/files/pkg/ > Reading http://releases.qt-project.org/pyside/ > Reading http://download.qt-project.org/official_releases/pyside/ > Best match: PySide 1.2.0 > Downloading > https://pypi.python.org/packages/source/P/PySide/PySide-1.2.0.tar.gz > #md5=a6f8ce6c0ab4ce858a6bdd10a3af3101 > Processing PySide-1.2.0.tar.gz > Writing > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\setup.cfg > Running PySide-1.2.0\setup.py -q bdist_egg --dist-dir > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\egg-dist-tmp-tj2oiv > Removing > c:\users\user_x\appdata\local\temp\easy_install-o3ol30\PySide-1.2.0\pyside_package > error: Setup script exited with error: Failed to find qmake. Please > specify thepath to qmake with --qmake parameter. > > Am I doing something wrong? > > Regarsd Robert > > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From matthew.woehlke at kitware.com Mon Jul 22 22:33:27 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Mon, 22 Jul 2013 16:33:27 -0400 Subject: [PySide] can't wrap subclass of QObject? Message-ID: I feel confident I am doing something dumb, but I can't figure out why, when trying to wrap any class that subclasses QObject (and/or has a Q_OBJECT?), I am getting this error: foo_wrapper.cpp: error: no ‘const QMetaObject* fooWrapper::metaObject() const’ member function declared in class ‘fooWrapper’ Most of the results trying to search the web for similar errors suggest an include path issue, however AFAICT I am passing the include directories for Qt to shiboken. My command line is (with some unrelated includes elided): /usr/bin/shiboken --generatorSet=shiboken --enable-parent-ctor-heuristic --enable-return-value-heuristic --enable-pyside-extensions /path/to/foo/build/foo/foo_all_sdk_headers.h --include-paths=/usr/include:[...elided....]:/usr/include:/usr/include/QtGui:/usr/include/QtXml:/usr/include/QtNetwork:/usr/include/QtCore:/path/to/foo/build:/path/to/foo/build/foo --typesystem-paths=/usr/share/PySide/typesystems --output-directory=/path/to/foo/build/foo /path/to/foo/build/foo/foo_typesystem.xml ...and my (simplified) typesystem XML: Any hints? -- Matthew From matthew.woehlke at kitware.com Mon Jul 22 23:15:13 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Mon, 22 Jul 2013 17:15:13 -0400 Subject: [PySide] can't wrap subclass of QObject? In-Reply-To: References: Message-ID: On 2013-07-22 16:33, Matthew Woehlke wrote: > I feel confident I am doing something dumb, but I can't figure out why, > when trying to wrap any class that subclasses QObject (and/or has a > Q_OBJECT?), I am getting this error: > > foo_wrapper.cpp: error: no ‘const QMetaObject* fooWrapper::metaObject() > const’ member function declared in class ‘fooWrapper’ > > Most of the results trying to search the web for similar errors suggest > an include path issue, however AFAICT I am passing the include > directories for Qt to shiboken. My command line is (with some unrelated > includes elided): > > /usr/bin/shiboken > --generatorSet=shiboken > --enable-parent-ctor-heuristic > --enable-return-value-heuristic > --enable-pyside-extensions /path/to/foo/build/foo/foo_all_sdk_headers.h > > --include-paths=/usr/include:[...elided....]:/usr/include:/usr/include/QtGui:/usr/include/QtXml:/usr/include/QtNetwork:/usr/include/QtCore:/path/to/foo/build:/path/to/foo/build/foo > --typesystem-paths=/usr/share/PySide/typesystems > --output-directory=/path/to/foo/build/foo > /path/to/foo/build/foo/foo_typesystem.xml > > > ...and my (simplified) typesystem XML: > > > > > > > > > Any hints? Possible lead... I am also seeing these warnings: type 'QMetaObject' is specified in typesystem, but not defined. This could potentially lead to compilation errors. signature 'metaObject()const' for function modification in 'QObject' not found. Possible candidates: ...and lots of others, but those look most relevant. -- Matthew From jpe at wingware.com Mon Jul 22 23:35:04 2013 From: jpe at wingware.com (John Ehresman) Date: Mon, 22 Jul 2013 17:35:04 -0400 Subject: [PySide] can't wrap subclass of QObject? In-Reply-To: References: Message-ID: <51EDA588.706@wingware.com> On 7/22/13 4:33 PM, Matthew Woehlke wrote: > Most of the results trying to search the web for similar errors suggest > an include path issue, however AFAICT I am passing the include > directories for Qt to shiboken. My command line is (with some unrelated > includes elided): > > /usr/bin/shiboken > --generatorSet=shiboken > --enable-parent-ctor-heuristic > --enable-return-value-heuristic > --enable-pyside-extensions /path/to/foo/build/foo/foo_all_sdk_headers.h > > --include-paths=/usr/include:[...elided....]:/usr/include:/usr/include/QtGui:/usr/include/QtXml:/usr/include/QtNetwork:/usr/include/QtCore:/path/to/foo/build:/path/to/foo/build/foo > --typesystem-paths=/usr/share/PySide/typesystems > --output-directory=/path/to/foo/build/foo > /path/to/foo/build/foo/foo_typesystem.xml Do you have a global.h ? The command line I use for a simple extension with the paths left out is: shiboken --generator-set=shiboken build/global.h --avoid-protected-hack --enable-pyside-extensions --include-paths=... --typesystem-paths=...--output-directory=build typesystem_WGuiCppImpl.xml Cheers, John From matthew.woehlke at kitware.com Mon Jul 22 23:47:30 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Mon, 22 Jul 2013 17:47:30 -0400 Subject: [PySide] can't wrap subclass of QObject? In-Reply-To: <51EDA588.706@wingware.com> References: <51EDA588.706@wingware.com> Message-ID: On 2013-07-22 17:35, John Ehresman wrote: > On 7/22/13 4:33 PM, Matthew Woehlke wrote: >> Most of the results trying to search the web for similar errors suggest >> an include path issue, however AFAICT I am passing the include >> directories for Qt to shiboken. My command line is (with some unrelated >> includes elided): >> >> /usr/bin/shiboken >> --generatorSet=shiboken >> --enable-parent-ctor-heuristic >> --enable-return-value-heuristic >> --enable-pyside-extensions /path/to/foo/build/foo/foo_all_sdk_headers.h >> >> --include-paths=/usr/include:[...elided....]:/usr/include:/usr/include/QtGui:/usr/include/QtXml:/usr/include/QtNetwork:/usr/include/QtCore:/path/to/foo/build:/path/to/foo/build/foo >> --typesystem-paths=/usr/share/PySide/typesystems >> --output-directory=/path/to/foo/build/foo >> /path/to/foo/build/foo/foo_typesystem.xml > > Do you have a global.h ? The command line I use for a simple extension > with the paths left out is: > > shiboken --generator-set=shiboken build/global.h --avoid-protected-hack > --enable-pyside-extensions --include-paths=... > --typesystem-paths=...--output-directory=build typesystem_WGuiCppImpl.xml Yes, that's what the '_all_sdk_headers' is. (Sorry, it should have been on a new line.) That said... I think I figured out the problem. From what I can tell, '#include ' doesn't really work for shiboken; one must '#include ' instead. -- Matthew From matthew.woehlke at kitware.com Tue Jul 23 01:32:10 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Mon, 22 Jul 2013 19:32:10 -0400 Subject: [PySide] manipulating >10 args? Message-ID: I'm trying to wrap the following method: MyDoubleInputDialog::getDouble(QWidget*, QString, QString, double, double, double, QString, int, double, bool*, QFlags<Qt::WindowType>) ...and I'm running into problems. Essentially I need to twist this a la (e.g.) QString::toInt, which seems like it should be straight-forward: bool ok_; %BEGIN_ALLOW_THREADS %RETURN_TYPE retval_ = %CPPSELF.%FUNCTION_NAME(%1, %2, %3, %4, %5, %6, %7, %8, %9, &ok_, %11); %END_ALLOW_THREADS ...however I'm getting a build error regarding "cppArg01". AFAICT what is happening is in "%11", the "%1" is being replaced, leaving a stray '1' and causing the eleventh argument to not be converted at all. Is there a technique I am missing for how this is supposed to work, or is handling of '%#' not working for # >= 10? I was able to work around this by patching shiboken: diff --git a/generator/shiboken/shibokengenerator.cpp b/generator/shiboken/shibokengenerator.cpp index 7de0d3a..30be324 100644 --- a/generator/shiboken/shibokengenerator.cpp +++ b/generator/shiboken/shibokengenerator.cpp @@ -1659,7 +1659,7 @@ void ShibokenGenerator::writeCodeSnips(QTextStream& s, if (type->isReference() || isPointer(type)) code.replace(QString("%%1.").arg(idx), QString("%1->").arg(replacement)); } - code.replace(QString("%%1").arg(idx), pair.second); + code.replace(QRegExp(QString("%%1\\b").arg(idx)), pair.second); } if (language == TypeSystem::NativeCode) { ...but I'd appreciate feedback if I am missing something, or if the above looks like a reasonable fix, or if there is a better way. -- Matthew From dennis.victorovich at gmail.com Tue Jul 23 13:01:54 2013 From: dennis.victorovich at gmail.com (Dennis Victorovich) Date: Tue, 23 Jul 2013 15:01:54 +0400 Subject: [PySide] manipulating >10 args? In-Reply-To: References: Message-ID: I'm not sure if its possible to pass more than 10 arguments into call. 10 arguments constraints may come from QMetaObject::Invoke() limitations ( http://qt-project.org/doc/qt-4.8/qmetaobject.html#invokeMethod) Take a look at cpp bindings code generated from your template. Shiboken may use QMetaObject::Invoke() 2013/7/23 Matthew Woehlke > I'm trying to wrap the following method: > > MyDoubleInputDialog::getDouble(QWidget*, QString, QString, double, > double, double, QString, int, double, > bool*, QFlags<Qt::WindowType>) > > ...and I'm running into problems. Essentially I need to twist this a la > (e.g.) QString::toInt, which seems like it should be straight-forward: > > > > > > > > > bool ok_; > %BEGIN_ALLOW_THREADS > %RETURN_TYPE retval_ = > %CPPSELF.%FUNCTION_NAME(%1, %2, %3, %4, %5, %6, > %7, %8, %9, &ok_, %11); > %END_ALLOW_THREADS > > > > > ...however I'm getting a build error regarding "cppArg01". AFAICT what > is happening is in "%11", the "%1" is being replaced, leaving a stray > '1' and causing the eleventh argument to not be converted at all. > > Is there a technique I am missing for how this is supposed to work, or > is handling of '%#' not working for # >= 10? > > > I was able to work around this by patching shiboken: > > diff --git a/generator/shiboken/shibokengenerator.cpp > b/generator/shiboken/shibokengenerator.cpp > index 7de0d3a..30be324 100644 > --- a/generator/shiboken/shibokengenerator.cpp > +++ b/generator/shiboken/shibokengenerator.cpp > @@ -1659,7 +1659,7 @@ void ShibokenGenerator::writeCodeSnips(QTextStream& > s, > if (type->isReference() || isPointer(type)) > code.replace(QString("%%1.").arg(idx), > QString("%1->").arg(replacement)); > } > - code.replace(QString("%%1").arg(idx), pair.second); > + code.replace(QRegExp(QString("%%1\\b").arg(idx)), pair.second); > } > > if (language == TypeSystem::NativeCode) { > > > ...but I'd appreciate feedback if I am missing something, or if the > above looks like a reasonable fix, or if there is a better way. > > -- > Matthew > > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From jpe at wingware.com Tue Jul 23 17:15:32 2013 From: jpe at wingware.com (John Ehresman) Date: Tue, 23 Jul 2013 11:15:32 -0400 Subject: [PySide] New pyside-dev list Message-ID: <51EE9E14.9060601@wingware.com> Hi, We've created a pyside-dev list at pyside-dev at googlegroups.com for discussions of the development of pyside. Discussions of how to use pyside should remain on this list. Anyone is free to join pyside-dev and archives are available at https://groups.google.com/forum/#!forum/pyside-dev Cheers, John From Fun_Extra_300 at gmx.net Tue Jul 23 18:08:55 2013 From: Fun_Extra_300 at gmx.net (Robert Schilling) Date: Tue, 23 Jul 2013 18:08:55 +0200 (CEST) Subject: [PySide] PySide calls wrong emit() inside class Message-ID: Hi @ all, Consider following example: from PySide import QtCore from PySide.QtCore import QObject class Foo(QObject): sig = QtCore.Signal(str) def __init__(self): QObject.__init__(self) def emit(self): self.sig.emit("1") x = Foo() x.emit() This example doesn't work. When calling self.sig.emit("1") actually Foo.emit() is invoked. Testing this example works with PyQt (that's what I'm migrating from). Am I doing something wrong? How can I avoid this? Regards Robert From nathanjsmith at gmail.com Tue Jul 23 18:31:28 2013 From: nathanjsmith at gmail.com (Nathan Smith) Date: Tue, 23 Jul 2013 11:31:28 -0500 Subject: [PySide] PySide calls wrong emit() inside class In-Reply-To: References: Message-ID: Hi Robert, You've created your signal at the class level, so it is shared by all instances of Foo. What you want is the signal at the object level, so each instance holds its own signal. Do so by creating the signal in the __init__ method. def __init__(self): QObject.__init__(self) self.sig = QtCore.signal(str) Nathan On Jul 23, 2013 11:09 AM, "Robert Schilling" wrote: > Hi @ all, > > Consider following example: > > from PySide import QtCore > from PySide.QtCore import QObject > > class Foo(QObject): > > sig = QtCore.Signal(str) > > def __init__(self): > QObject.__init__(self) > > def emit(self): > self.sig.emit("1") > > x = Foo() > x.emit() > > This example doesn't work. When calling self.sig.emit("1") actually > Foo.emit() is invoked. > Testing this example works with PyQt (that's what I'm migrating from). > > Am I doing something wrong? How can I avoid this? > > Regards Robert > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside > -------------- next part -------------- An HTML attachment was scrubbed... URL: From Fun_Extra_300 at gmx.net Tue Jul 23 18:36:20 2013 From: Fun_Extra_300 at gmx.net (Robert Schilling) Date: Tue, 23 Jul 2013 18:36:20 +0200 (CEST) Subject: [PySide] PySide calls wrong emit() inside class In-Reply-To: References: , Message-ID: Hi Nathan, Thanks for the answer, but this doesn't work either: from PySide import QtCore from PySide.QtCore import QObject class Foo(QObject): def __init__(self): QObject.__init__(self) self.sig = QtCore.Signal(str) def emit(self): self.sig.emit("1") x = Foo() x.emit() This triggers the following error: self.sig.emit("1") AttributeError: 'PySide.QtCore.Signal' object has no attribute 'emit' Having the classwide signal, this is from the type ySide.QtCore.Signalnstance. Robert     Gesendet: Dienstag, 23. Juli 2013 um 18:31 Uhr Von: "Nathan Smith" An: "Robert Schilling" Cc: pyside at qt-project.org Betreff: Re: [PySide] PySide calls wrong emit() inside class Hi Robert, You've created your signal at the class level, so it is shared by all instances of Foo. What you want is the signal at the object level, so each instance holds its own signal. Do so by creating the signal in the __init__ method. def __init__(self):    QObject.__init__(self)    self.sig = QtCore.signal(str) Nathan On Jul 23, 2013 11:09 AM, "Robert Schilling" wrote:Hi @ all, Consider following example: from PySide import QtCore from PySide.QtCore import QObject class Foo(QObject):     sig = QtCore.Signal(str)     def __init__(self):         QObject.__init__(self)     def emit(self):         self.sig.emit("1") x = Foo() x.emit() This example doesn't work. When calling self.sig.emit("1") actually Foo.emit() is invoked. Testing this example works with PyQt (that's what I'm migrating from). Am I doing something wrong? How can I avoid this? Regards Robert _______________________________________________ PySide mailing list PySide at qt-project.org[PySide at qt-project.org] http://lists.qt-project.org/mailman/listinfo/pyside From matthew.woehlke at kitware.com Tue Jul 23 18:45:56 2013 From: matthew.woehlke at kitware.com (Matthew Woehlke) Date: Tue, 23 Jul 2013 12:45:56 -0400 Subject: [PySide] manipulating >10 args? In-Reply-To: References: Message-ID: On 2013-07-23 07:01, Dennis Victorovich wrote: > I'm not sure if its possible to pass more than 10 arguments into call. > 10 arguments constraints may come from QMetaObject::Invoke() limitations ( > http://qt-project.org/doc/qt-4.8/qmetaobject.html#invokeMethod) > Take a look at cpp bindings code generated from your template. Shiboken may > use QMetaObject::Invoke() I'd expect it to be calling the code I gave it in : %RETURN_TYPE retval_ = %CPPSELF.%FUNCTION_NAME(%1, %2, %3, %4, %5, %6, %7, %8, %9, &ok_, %11); ...and it seems to be doing so: double retval_ = qtDoubleInputDialog::getDouble(cppArg0, cppArg1, cppArg2, cppArg3, cppArg4, cppArg5, cppArg6, cppArg7, cppArg8, &ok_, cppArg9); (line re-wrapped for readability in e-mail) There is no 'invoke' (case-insensitive search) anywhere in the generated code for this class. There may be cases where that is a real limit (FWIW I'll also point out that this particular method is a static class method and - as such, AFAIK - couldn't be called with QMetaObject::invokeMethod anyway), but in general I would say shiboken ought to handle more than nine args. And it seems the only problem is that the '%#' substitution doesn't check for whole words (or operate from highest-number to lowest), such that '%10', etc. ends up replaced as '%1' with a leftover digit. So an easy and obvious fix is to ensure that '%1' is only replaced if it is not followed by another digit, for which using a regex ending in '\b' suffices. -- Matthew From jpe at wingware.com Tue Jul 23 21:06:30 2013 From: jpe at wingware.com (John Ehresman) Date: Tue, 23 Jul 2013 15:06:30 -0400 Subject: [PySide] PySide calls wrong emit() inside class In-Reply-To: References: , Message-ID: <51EED436.9060100@wingware.com> On 7/23/13 12:36 PM, Robert Schilling wrote: > Hi Nathan, > > Thanks for the answer, but this doesn't work either: > > from PySide import QtCore > from PySide.QtCore import QObject > > class Foo(QObject): > def __init__(self): > QObject.__init__(self) > self.sig = QtCore.Signal(str) > > def emit(self): > self.sig.emit("1") emit() is defined by QObject and you're overriding it here. I suspect this is not what you intend and you probably want to use a different name. Cheers, John From lucericml at gmail.com Wed Jul 24 18:08:43 2013 From: lucericml at gmail.com (Luc-Eric Rousseau) Date: Wed, 24 Jul 2013 12:08:43 -0400 Subject: [PySide] new crash in pyside 1.1.2 in signalInstanceConnect() Message-ID: Hello, my team has filed this bug: https://bugreports.qt-project.org/browse/PYSIDE-179 From jmohler at gamry.com Wed Jul 24 21:19:14 2013 From: jmohler at gamry.com (Joel B. Mohler) Date: Wed, 24 Jul 2013 15:19:14 -0400 Subject: [PySide] new crash in pyside 1.1.2 in signalInstanceConnect() In-Reply-To: References: Message-ID: <51F028B2.3080507@gamry.com> On 7/24/2013 12:08 PM, Luc-Eric Rousseau wrote: > Hello, my team has filed this bug: > https://bugreports.qt-project.org/browse/PYSIDE-179 You may have discovered this, but the easy work-around for this bug is to hold a reference to the selection model returned by catView.selectionModel(). What I've done is simply do something like this: self.selmodel = catView.selectionModel() self.selmodel.selectionChanged.connect(selectionChanged) But, yes, it'd be nice if this was fixed ... Joel From chigga101 at gmail.com Tue Jul 30 18:30:32 2013 From: chigga101 at gmail.com (Matthew Ngaha) Date: Tue, 30 Jul 2013 17:30:32 +0100 Subject: [PySide] mobile question Message-ID: hey guys i slowed down a bit in my learning of Pyside but i noticed i focused a lot on qml. The problem i have is everytime i see an instructional qml video, to get an application on a mobile device, they always use QtCreator. Always. It's left me lost because i know this isn't available to Python, and having never done mobile apps i want to know is it possible to get a qml or even pyside app on a mobile device without QtCreator, and how? Is it a hard process or still very simple like it would be with C++? From martin.kolman at gmail.com Tue Jul 30 18:51:31 2013 From: martin.kolman at gmail.com (Martin Kolman) Date: Tue, 30 Jul 2013 18:51:31 +0200 Subject: [PySide] mobile question In-Reply-To: References: Message-ID: <51F7EF13.8060703@gmail.com> 30.7.2013 18:30, Matthew Ngaha: > hey guys i slowed down a bit in my learning of Pyside but i noticed i > focused a lot on qml. The problem i have is everytime i see an > instructional qml video, to get an application on a mobile device, > they always use QtCreator. Always. It's left me lost because i know > this isn't available to Python, and having never done mobile apps i > want to know is it possible to get a qml or even pyside app on a > mobile device without QtCreator, and how? Is it a hard process or > still very simple like it would be with C++? > _______________________________________________ > PySide mailing list > PySide at qt-project.org > http://lists.qt-project.org/mailman/listinfo/pyside Depends on the platform: * on Maemo (Nokia N900) and MeeGo 1.2 Harmattan (Nokia N9), you just have to manually make a Debian package for your application * for Android, you can actually use the Qt Creator from the Necessitas SDK to deploy & debug your PySide application, see: http://qt-project.org/wiki/PySide_for_Android_guide https://github.com/raaron/pydroid * for BB10, you just have to roll your own BAR and you are good to go (not much documentation for this at the moment though) * for Nemo Mobile, you just have to make a RPM package manually To just run the application, you can usually get around a full package-deploy cycle by just deploying the application once and then replacing the Python or QML part by some rsync/scp/adb or similar. Enables quite a rapid debug cycle. :) BTW, my PySide-using applications - modRana[1] and Mieru[2] work & are packaged for all of these platforms, so if you have any questions, don't hesitate to ask. :) [1] https://github.com/M4rtinK/modrana [2] https://github.com/M4rtinK/Mieru Best wishes Martin Kolman From khertan at khertan.net Tue Jul 30 19:19:03 2013 From: khertan at khertan.net (khertan at khertan.net) Date: Tue, 30 Jul 2013 17:19:03 +0000 Subject: [PySide] mobile question In-Reply-To: References: Message-ID: It s depends on what is available on the mobile platform ... PySide isn't available for Qt5 so Ubuntu bidule and Jolla Sailfish os can't run PySide. PyQt could probably be used on Ubuntu and Jolla, an other is pyotherside : http://github.com/khertan/pyotherside You can use pyside on MeeGo Harmattan no need of qtcreator ... u can even code on a Nokia n9 : some examples apps : http://khertan/KhtSimpleText, http://khertan/BitPurse ... Regards -- Benoît HERVIER - http://khertan.netLe 30/07/13 18:30 Matthew Ngaha a écrit : hey guys i slowed down a bit in my learning of Pyside but i noticed i focused a lot on qml. The problem i have is everytime i see an instructional qml video, to get an application on a mobile device, they always use QtCreator. Always. It's left me lost because i know this isn't available to Python, and having never done mobile apps i want to know is it possible to get a qml or even pyside app on a mobile device without QtCreator, and how? Is it a hard process or still very simple like it would be with C++? _______________________________________________ PySide mailing list PySide at qt-project.org http://lists.qt-project.org/mailman/listinfo/pyside From chigga101 at gmail.com Tue Jul 30 20:34:58 2013 From: chigga101 at gmail.com (Matthew Ngaha) Date: Tue, 30 Jul 2013 19:34:58 +0100 Subject: [PySide] mobile question In-Reply-To: References: Message-ID: Hi khertan, The links to your examples don't work. That pyotherside looks very interesting. Can you tell me if it is possible for pyqt to get qml apps on mobile without QtCreator? Also Please do you know the status of PySide? why hasn't it been updated to have Qt5 support? development still ongoing? From khertan at khertan.net Tue Jul 30 21:16:14 2013 From: khertan at khertan.net (khertan at khertan.net) Date: Tue, 30 Jul 2013 19:16:14 +0000 Subject: [PySide] mobile question In-Reply-To: References: Message-ID: oh sorry, some words missing, and link are http://khertan.net/BitPurse http://khertan.net/KhtSimpleText even for writing qt c++ apps you can do it without qtcreator... And why it's not available yet for Qt5 ? ... not enough people to give blood ... -- Benoît HERVIER - http://khertan.netLe 30/07/13 20:34 Matthew Ngaha a écrit : Hi khertan, The links to your examples don't work. That pyotherside looks very interesting. Can you tell me if it is possible for pyqt to get qml apps on mobile without QtCreator? Also Please do you know the status of PySide? why hasn't it been updated to have Qt5 support? development still ongoing?