General and Desktop

@Kent-Dorfman I'm setting the style app.setStyle(QStyleFactory.create('windowsvista')) like this, and since I'm using a Windows only audio control library, I don't really care for other OS'es

@JonB Thanks for the hint, but im in no hurry right now. :D Have a nice day.

Hi and welcome to devnet, Something is not clear: are you creating a full custom widget ? It seems so but since Qt provides the base classes to create your own custom views within the model view set of classes, it would be good to clarify that first.

So i finally found out what the problem was and fixed it... it seems i had given specific size settings to some component via a stylesheet, and some components did not have some/all properties defined in the stylesheet. Providing the correct info via the stylesheet appears to have resolved the situation. Eg. for the radio buttons i now have... QRadioButton::indicator { width: 14px; height: 14px; } in my style sheet, and it seems to properly respect the scaling when i have both types of scaling in use.

@JonB yeah, I've seen this before. If I'm reading this case correctly, the variable s2 is no longer needed by line 115, and has either been trashed or no longer properly maintained by the debugger.

@JonB Definitely is, thanks again for the help Jon

Hi and welcome to devnet, AFAIK, it literally means that the file path you passed points to something that does not exist. Did you check the path for correctness ? Where does it point to ? What kind of file is that ?

Thanks for the reply! Yes, I agree this really seems to be a tvirt.vir issue, not Qt. That's why I wrote a standalone C program (outside Qt) to test the pipe communication. I’ve already tried: Sending the message DO#0#1 to \\.\pipe\\tubotest Triggering the EventoEnviar event after writing Still, nothing changes on the simulator side, and the input value shown in my DAQ console remains 0. I do have some documentation, and it says tvirt.vir listens for specific commands and events, but I’ll try to confirm exactly what sequence it expects. Thanks again. I’ll keep testing with the C version for now.

@JulsPower That does look more reasonable, i wasn't sure what the MS docs were telling you to do. Which implies you were in the Another possibility is that the issue has nothing to do with Qt or deployment but rather is a "bug" which only happens to show up in certain environments or circumstances. Such as a memory access issue. I had a quick glance at your WTSEnumerateSessionsExA() and it seemed OK to me, but you might want to try it standalone without any Qt stuff to make sure that works across environments. situation! The Qt and deployment changes simply showed or masked the problem. But you have still learnt what you should do to e.g. deploy to another machine, or be sure it works on yours outside of the Creator development environment which should be useful :)

@aha_1980 said in How to switch off specific warnings in Qt Creator: @DiBosco Have you already seen https://stackoverflow.com/questions/22129383/removing-unused-parameters-warning-in-qtcreator ? For your case, it would be: QMAKE_CXXFLAGS_WARN_ON += -Wno-deprecated-copy Regards Ah, yes I had seen that. The English in the most upticked reply just made me go "What?". Just could not understand what they were saying, but I had tried: QMAKE_CXXFLAGS_WARN_ON += -Wdeprecated-copy Thinking why on earth would putting the warning on make it disappear, but out of desperation I gave it a go. Had I realised it needed -Wno-deprecated-copy, not -Wdeprecated-copy I'd've been fine hours ago. Many thanks for that, now doing just what I want. Also @JonB Thanks for your help too.

@hskoglund said in Qt6 and QT_ENABLE_HIGHDPI_SCALING: If you right-click on the .exe file, select Properties and in the Compability tab click the "Change high DPI settings" button, does any setting in that dialog box help the faulty scaling after QApplication is instantiated? Thanks for the suggestion but I think those settings effectively override the manifest (so startup only) and indeed I can change the setting on an exe so that it "supports" High DPI even when it doesn't and start it up and you can see scaling issues. Reverting it while it's running makes no difference. However, this is now moot as I decided to suck it up and change the legacy code to check for and support the system scaling. We have a fair number of these MFC apps and on some this will be a lot more work than others. This one is not too bad as the main client window is written using our Direct2D library which already supports such scaling. As expected Windows common controls support this out of the box and weirdly all our MFC dialogs also display scaled with the only apparent issue being the font text display appearing "thin" and somewhat faded. Not enough of a problem to care about.

I think it is necessary to port the Qt5 service library to Qt6. I haven't did it because the qt6 already built framework can't run on Windows 2016 machines and I have a lot of them. Qt6 requires Windows 10. So, up to now, I'm using Qt5.

@jsulm said in System tray crash in macOS 16: how is your question related to this thread? Forked now :)

@Axel-Spoerl, I only wanted to discuss it with them, but you've all answered everything I wanted to know. (I don't mean I was speaking on their behalf.) Thanks, all!

Just see here and here.

@Perdrix said in Size of QtWebEngine: Unfortunately litehtml won't work for me - I need function it doesn't provide :( That's pretty much the tradeoff. Chromium has All The Things you might need, but that means it has the disk usage of All The Things.

@Perdrix said in QProcess::terminate() or QProcess::close(): All well and good but that presumes that the process in question accepts input asking it nicely to close (probably on re-directed stdin). In this case the help display process doesn't support that. I've now changed to use start() instead of startDetached() Actually that's not entirely true. For a normal command line program it would terminate on its own, such as "ls, ps ax, df,...whatever", but in your case it's a little different. The POSIX way to tell a child to die gracefully is to send it a kill signal, which is different from using close/terminate. kill(pid, SIGTERM) kills a process, and can be trapped by the child to do cleanup. But in all fairness the QProcess::close() probably does that implicitly, whereas QProcess::terminate() is not "nice" about it.

@bogong said in Add resource file (*.qrc) to static library. How?: How to add *.qrc files to library and how to use it in the main project? You would use qt_add_resources() for *.qrc files: https://doc.qt.io/qt-6/qt-add-resources.html However, if your *.qml files are inside a *.qrc then you won't get the full benefits of qt_add_qml_module() (such as optimization via the Qt Quick Compiler: https://www.qt.io/blog/the-numbers-performance-benefits-of-the-new-qt-quick-compiler). It is better to move all your files out of the *.qrc, and directly into the qt_add_qml_module(): qt_add_qml_module(${A_NAME_TARGET} URI Library_v1 QML_FILES ATestQML.qml BTestQML.qml RESOURCES myimage.png myotherimage.png )

@EduardoQtDev said in CMake problem with Qt 6.10.0: Just in case someone wants to open a bug report. https://bugreports.qt.io/browse/QTBUG-137577 - Known, and already fixed it seems. You'll need to wait for beta2, or build from Git.