General and Desktop

You would need to escape literal double-quotes if they are also present within the regular expressions. Using a JSON package/library, like QJSonDocument, to write the file is the easiest way to ensure that escaping happens where it should (and is reversed on read). Another observation: your email address regex is not allowing for a lot of options. This, for example, is a valid email address, "someone+something@example.com". See this Stackoverflow article for more.

@memyselfandqt said in mysterious warning: I know I must modify this file in design. So, I never edit this ui file. But you seem to have deleted something (maybe not intentionally) See https://forum.qt.io/post/566146

I got it going. Answer is here https://stackoverflow.com/questions/78017906/need-xpath-2-0-c-for-mingw/78028760#78028760

@apaczenko1993 Sorry. Qt 5.15.2 seems not having pdf module as default in the installer. I am using 5.15.3 on Ubuntu 22.04 and pdf module can be installed. Qt 6.6.2 has it as default.

@a_coder As both other responders said, you need a vertical spacer above a widget if you want to push it down.

Hi, I faced the same problem, failed to load OPC UA plugin: qt.opcua: Failed to load OPC UA plugin: "open62541" qt.opcua: Available plugins: QList("open62541") Could not create client I solved by putting libcrypto-3-x64.dll in the output build directory. Open62541_backend.dll depends on libcrypto-3-x64.dll, you can check it by using Dependency Walker tool. My env: Qt 6.6.0, MSVC2019 64bit.

@jsulm said in Drawing QRubberBands with variabe size, but fixed aspect ratio: how to do it?: rubberBand->setGeometry(QRect(origin, event->pos()).normalized()); Thanks. It is indeed as easy as you indicated! Regards Bertwim

@JoeCFD Yes you are right. I was stay on version 5.12.12. I have just try with version 6.6.1 and the control works has expected. Very surprise this kind of "bug"/"strange behavior" has stay on version 5.12 which was an LTS. Anyway thanks a lot for your answer

@Slingy Hello Slingy, i am experiencing exact same behaviour. i know that the qt bluetooth api is not entirely compatible with mingw in qt 5.14. but switching to msvc compiler is not an option for me. Did you get any further on this?

@Jammin44fm Thanks for everyones input I've traced the issue to a problem with process input, and i'm going to create a more specific issue to address that.

It should also be possible to use Windows application and DLL manifest entries to achieve DLL redirection during program start. This is a dark art AFAICT.

Hi @MrHandSOme, I calculate by putting qtime in the upper and lower parts. QTime has some overhead (probably very little, but still...). For interest, try this: const quint64 start = QDateTime::currentMSecsSinceEpoch(); QJsonObject rootObj; qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "rootObj"; rootObj[QStringLiteral("a1")]=QStringLiteral("a1obj"); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "a1obj"; rootObj[QStringLiteral("a2")]=QStringLiteral("a2obj"); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "a2obj"; rootObj[QStringLiteral("a3")]=QStringLiteral("a3obj"); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "a3obj"; rootObj[QStringLiteral("a4")]=QStringLiteral("a4obj"); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "a4obj"; rootObj[QStringLiteral("a5")]=QStringLiteral("a5obj"); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "a5obj"; QJsonDocument newDocument(rootObj); qDebug() << (QDateTime::currentMSecsSinceEpoch()-start) << "newDocument"; For me (on Linux though, I don't have Windows handy just at the moment), it shows all of that code executing within a single millisecond, as expected. My output: 0 rootObj 1 a1obj 1 a2obj 1 a3obj 1 a4obj 1 a5obj 1 newDocument If that code is slow for you, show us the output. If it's not slow, then try replacing the various QStringLiteral()'s with your versions one at a time and see if that reveals anything. Cheers.

I have found the answer, or at least what I think is the answer. Providing it here in case anyone else ever has this problem, I don't think it was very obviousm It seems to be to use QLabel.setBuddy(arg) on the label object passing it the widget to be labelled. Making this call gets NVDA to correctly identify the label for the widget.

@McLion Would you currently prefer this approach over Qt D-Bus for it's simplicity?

@Joachim-W said in Style-sheet pseudo-state for selected item in QListView?: As I learned from https://forum.qt.io/topic/154635, and as confirmed by the API Docs (https://doc.qt.io/qt-6/qabstractitemview.html#currentIndex), there is also a "current" item in a QListView. As @JonB said, you are mixing up code "logic" and stylesheet states... They have no direct relation to each other. If I create my own item view and add a function called blaa(), which returns a struct named Blaa, it does not mean you have a property or style selector named MyItemView:blaa to apply any stylesheet to.