General and Desktop

As advised on the #qt freenode channel, this is due to the way in which I was declaring circle widgets when calling it from window.cpp, I had missed off the pointer definition (*) when instantiating it. CircleWidget* circleWidgets = new CircleWidget I then received an error about vtables, so need to work on the Q_OBJECT macro in the header files.

Yes, my mistake using * (pointer) like parameter in the signal. I just changed to this: @ SearchRequest::SearchRequest(QObject *parent) : QObject(parent) { QNetworkAccessManager *manager = new QNetworkAccessManager(this); QUrl url("url here..."); QNetworkRequest request(url); request.setHeader(QNetworkRequest::ContentTypeHeader, "application/x-www-form-urlencoded"); QUrl params; params.addQueryItem("param1", "..."); params.addQueryItem("param2", "..."); params.addQueryItem("param3", "..."); connect(manager, SIGNAL(finished(QNetworkReply *)), this, SLOT(rfinished(QNetworkReply *))); QNetworkReply *reply = manager->post(request, params.encodedQuery()); } void SearchRequest::rfinished(QNetworkReply *reply) { QByteArray data = reply->readAll(); QString str(data); emit test(data); } signals: void test(QString text); @

Please be sure and edit the title of the original post to add [Solved]. Thanks!

[quote author="wssddc" date="1356650243"]I don't extend QTableWidget. The horizontal header is part of it. I just do @ connect(ui->tableWidget->horizontalHeader(), SIGNAL(sectionClicked(int)), this, SLOT(slot_table_clicked(int))); @ and slot_table_clicked gets the column number that was clicked. [/quote] I hadn't realised I could connect signals and slots in this way. That solved my problem perfectly, thanks!

I too facing issue with debugging Qt on Windows. Let me tell you workaround [probably this may not fix your problem, but can be alternative to debug your application]: Note: This is valid only if you are compiling Qt application with MSVC2008 (Qt SDK) or MSVC2010 (Qt SDK) in your Qt Creator[i.e basically if you are compiling with nmake] Successfully build your application with Qt creator Open "Visual Studio Command Prompt (2008)" or "Visual Studio Command Prompt (2010)" from start menu > All Programs > Microsoft Visual Studio 20xx > Visual Studio Tools Navigate to location where your exe is built, for example cd c:\MyApp\debug\ in command prompt. run command devenv MyApp.exe This should open Visual Studio 20xx solution explorer Now in Visual Studio 20xx go to File > Open ; and open all your source files (*.cpp and .h files for MyApp) Keep break-points where ever you wish and press F5. You must be able to debug your application now on Visual Studio environment.

The leak is caused by @settings.autoLoadImages: false@ setting it to true makes qml work fine

OK. It is by the way senseless to put a socket into a thread! I am now using the thread-worker approach and it is working perfectly. If a signal is connected to a QObject (the worker) which is in another thread (one can use QObject::moveToThread()) the signal will be appended to a queue. The thread worker will receive the signals when the previos signal has been proceeded completely!

I tested this problem with a QTextEdit control, QTextEdit have the same problem - it can't display the complex Chinese character.

Thank you Lukas for the help. As mentioned above, QT3 compatible members will not be working in QT5. We should try to modify the usage of deprecated members as much as possible. If in case there is no choice left then only we can use the define in project (.pro) file (first statement possible) as @DEFINES += QT_DISABLE_DEPRECATED_BEFORE=0x000000@

The post is very good.

HW acceleration is disabled. That is, the option "3D acceleration" under "Display" - "Extended features" is disabled. It was enabled at first, and I got a ton more errors. Edit: Here are the errors I get with HW acceleration enabled: @OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo OpenGL Warning: XGetVisualInfo returned 0 visuals for 0xa21e2a0 OpenGL Warning: Retry with 0x8002 returned 0 visuals OpenGL Warning: glXGetFBConfigAttrib for 0xa21e2a0, failed to get XVisualInfo@ Edit 2: I should stress that the program paints correctly either way - with HW acceleration enabled or disabled. Also, the Linux in the VirtualBox runs pretty slowly (especially new windows are real slow to appear) - maybe there's a configuration problem. Edit 3: I tried the same VirtualBox on a better computer, it runs smoothly there, but the error messages are the same.

It turns out the line @ _mapView = new MapView(_parent); @ was clipping the view. By setting the parent to NULL in my test cases, I got the whole view saved in the QImage.

Fixed by: @doc.toString(0).replace(">\n<tspan", "><tspan")@ QDomDocument by default write each tag on new line, but by "SVG spec":http://www.w3.org/TR/SVGTiny12/text.html#TSpanElement <tspan> on new line equated to space.

I'm happy to hear that. Please add [Solved] to the beginning of this topic's subject. Happy coding!

Headers are private for a reason. If you choose to use them, you are doing it on your responsibility. There is no promise of any compatibility for private APIs, even between patch releases. Of course, in real life, most of the code will stay the same for long periods of time. There is nothing inherently wrong about it, you just need to recompile your project every time you update the libraries, and be very careful to use right version with your application, including user machines.