On 08.10.2015 22:09, Thiago Macieira wrote:
> On Thursday 08 October 2015 21:57:09 Igor Mironchik wrote:
>> Hi guys,
>>
>> Can anybody explain me, please. Is here memory leaks or no?
>>
>> Please look in the attachment...
> Yes, these look like real leaks:
>
> ==3533== 136 (16 direct, 120 indirect) bytes in 1 blocks are definitely lost
> in
> loss record 396 of 514
> ==3533== at 0x4C2C100: operator new(unsigned long) (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==3533== by 0x5EDF731: QQmlFileSelectorPrivate::QQmlFileSelectorPrivate()
> (in /home/igor/Qt/5.5/gcc_64/lib/libQt5Qml.so.5.5.0)
> ==3533== by 0x5EDF949: QQmlFileSelector::QQmlFileSelector(QQmlEngine*,
> QObject*) (in /home/igor/Qt/5.5/gcc_64/lib/libQt5Qml.so.5.5.0)
> ==3533== by 0x5ED8C26: QQmlApplicationEnginePrivate::init() (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Qml.so.5.5.0)
> ==3533== by 0x40272C: Application::Application(int, char**)
> (application.cpp:15)
> ==3533== by 0x402658: main (main.cpp:11)
>
> ==3533== 152 (112 direct, 40 indirect) bytes in 1 blocks are definitely lost
> in
> loss record 400 of 514
> ==3533== at 0x4C2C100: operator new(unsigned long) (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==3533== by 0x650EF21: QObject::QObject(QObject*) (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Core.so.5.5.0)
> ==3533== by 0x4F74A73: ??? (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
> ==3533== by 0x4F8340C: QSGRenderContext::initialize(QOpenGLContext*) (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
>
> ==3533== 224 (72 direct, 152 indirect) bytes in 1 blocks are definitely lost
> in
> loss record 414 of 514
> ==3533== at 0x4C2C100: operator new(unsigned long) (in
> /usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
> ==3533== by 0x4F7536B: ??? (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
> ==3533== by 0x4F8364B: QSGRenderContext::createTexture(QImage const&) const
> (in /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
> ==3533== by 0x4F2DF5B: ??? (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
> ==3533== by 0x4F8554B:
> QSGRenderContext::textureForFactory(QQuickTextureFactory*, QQuickWindow*) (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Quick.so.5.5.0)
>
> The backtrace of the first one goes all the way back to your Application
> constructor (which is missing the & for argc!!), so the leak might be in your
> code, not in Qt.
I look, look and look at this code, but I can't understand where leak
can be...
Application::Application(int & argc, char *argv[])
: QGuiApplication(argc, argv)
{
m_engine = new QQmlApplicationEngine(this);
auto context = m_engine->rootContext();
context->setContextProperty("appDir", applicationDirPath());
context->setContextProperty("App", this);
m_engine->load(QUrl("qrc:/MainWindow.qml"));
QObject *root = m_engine->rootObjects().at(0);
if(!root) {
qDebug() << "FAILED TO LOAD ROOT OBJECT";
::exit(ROOT_NOT_FOUND);
}
m_updateTimer = startTimer(1000);
}
>
> I'm assuming that all ??? that your log shows are not Qt frames and not of
> Qt's fault. In most cases, I'm probably right. There are some that I'd like
> more information on, like:
>
> ==3533== 42 bytes in 1 blocks are definitely lost in loss record 212 of 514
> ==3533== at 0x4C2BBA0: malloc (in /usr/lib/valgrind/vgpreload_memcheck-
> amd64-linux.so)
> ==3533== by 0x142E5F6E: ???
> ==3533== by 0x142C9314: ???
> ==3533== by 0xEA7D163: ???
> ==3533== by 0xEA7D518: ???
> ==3533== by 0xDCBAF59: ???
> ==3533== by 0xDCB1D11: ???
> ==3533== by 0xDC3D0DD: ???
> ==3533== by 0x54D46D6: QAccessible::setRootObject(QObject*) (in
> /home/igor/Qt/5.5/gcc_64/lib/libQt5Gui.so.5.5.0)
>
> Please recompile in debug mode and set the XCB plugin to no-unload (add to the
> linker flags: -Wl,-z,nodelete).
--
Best Regards,
Igor Mironchik.
_______________________________________________
Interest mailing list
Interest@qt-project.org
http://lists.qt-project.org/mailman/listinfo/interest
