General and Desktop

@Jammin44fm in case anyone is still using QtScript, I have an, AI assisted, experimental but functional port for Qt6. https://github.com/JulienMaille/qtscript-qt6 Feel free to test it and comment

Excellent ! Good luck :-)

@MarkoProg said in How does the tdesktop render tabs?: They don't use QML; they render everything from scratch using QWidgets, and I'm also interested in this approach. How exactly did they achieve this using QWidgets? The idea the same, regardless of whether you're using Qt Quick or Qt Widgets: A Model stores an overview of the chat data. One row in the Model represents one chat. A row would contain data like chat title, chat icon, latest timestamp, latest message, delivery status. A Delegate lays out the GUI components in its rectangle (e.g. icon on the left, timestamp of the top-right). A ListView (Qt Quick) or QListView/QListWidget (Qt Widgets) is used to display the "tabs". If the Model contains N rows (N chats), then the View creates N Delegates. The View takes data from each row and gives that data to corresponding Delegate of that row. The Delegate then writes the data into the GUI

How are you processing things currently ? Using multiple threads ? Involving the GPU ? What kind of processing are you doing ?

I made a bug report and added more details here https://qt-project.atlassian.net/browse/QTBUG-143646?focusedCommentId=2611561 but they haven't updated it yet, what do I do?

@DuBu your model will not update in your qml gui without Q_OBJECT macro in TableModel. I did not change anything in your code and it works on ubuntu .

Hi, Approach 1 is correct. MainWindow does not need to expose its internal structure in this case. In fact, it is pretty rare that an application main window exposes its internal state like for your solution #2. One other good point is creating your API client outside of MainWindow. This allows scenarios such as passing a fake API client for testing purpose if for any reason you cannot contact/spin up your backend server. This does not seem to be the case here but still it can be useful.

Thanks for the clarification. That explains the behavior. The limitation makes sense for my use case (Qt used as a plugin inside a non-Qt application). I had interpreted the documentation to mean that, as long as only one QApplication instance exists at a time, recreating it in a different thread would be supported. Knowing that QGuiApplication leaves behind internal objects tied to the first thread, making cross-thread recreation currently impossible, answers my question. I'll adjust the design accordingly. Thanks for the detailed explanation and the Qt bug reference.

# --- מציאת Qt6 לפני include(QtInstallPaths) --- find_package(QT NAMES Qt6 REQUIRED) find_package(Qt${QT_VERSION_MAJOR} REQUIRED COMPONENTS Core Widgets UiPlugin # נדרש לתוספי Designer — הסר אם לא רלוונטי Designer # נדרש לתוספי Designer — הסר אם לא רלוונטי )

It seems to be a bug. Reported 2 years ago. https://qt-project.atlassian.net/browse/QTBUG-127124

Hi, Which version of Qscintilla is that ? With which version of Qt ? Can you provide a minimal compilable example so people can reproduce the situation the same as you do ?

Maybe you can compare the cmake files from Qt for Qt6.11.0 and 6.11.1 to find a differerence which may affect the manifest generation. /edit: This looks promising: --- a/cmake/QtBaseGlobalTargets.cmake +++ b/cmake/QtBaseGlobalTargets.cmake @@ -459,6 +459,13 @@ elseif(WASM) "${QT_BUILD_DIR}/${INSTALL_LIBEXECDIR}/qt-wasmtestrunner.py" @ONLY) qt_install(PROGRAMS "${QT_BUILD_DIR}/${INSTALL_LIBEXECDIR}/qt-wasmtestrunner.py" DESTINATION "${INSTALL_LIBEXECDIR}") +elseif(WIN32) + qt_copy_or_install(FILES "cmake/windows/app.exe.manifest.in" + DESTINATION "${__GlobalConfig_install_dir}/windows" + ) + file(COPY "cmake/windows/app.exe.manifest.in" + DESTINATION "${__GlobalConfig_install_dir}/windows" + ) endif() # Install CI support files to libexec. --> https://codereview.qt-project.org/c/qt/qtbase/+/718374 If you find a solution which works for cl and clang-cl it would be nice.

Does your software have a style sheet set up somewhere?

The idea of the QValidator is to validate while you are typing/changing the value. If you just want to validate at the end (not while things are changed), then you can just write a function that goes over all the widgets. I remember a talk by an Adobe employee that was about something similar. AI tells me the speaker is Sean Parent and the library is Adam (a.k.a the Property Model Library or PML) which is part of the Adobe Source Libraries (ASL). Maybe you can find it and adapt some of their ideas. It is not explicitly written for Qt and might not work with it out of the box.

@jakubd How is this related to Qt? Did you debug to see where it crashes? What type of crash is that? Run in debugger and provide more information to answer above questions.

@SGaist Qt 6 and windows 11 It looks like what was missing was connect(scene(), &QGraphicsScene::sceneRectChanged, this, [this](QRectF rect) {this->setSceneRect(rect);}); Seems to work better, although not perfect.

I would say that Microsoft Word has already established what users expect when editing files together: You see all changes happening immediately together with a colored cursor for each user with a tag of their name. And those cursors move around live. The best thing is that you can follow where other users are currently editing the document and you'll most likely avoid editing the same thing. This usually requires for your documents to live in a "cloud" (even if the cloud is your own private cloud at home) and a server that at least tracks all concurrent users of the document and tells each user about the others. The best example of what it should have been like is Google Wave (unfortunately, this project has been shut down after only a few years). It was based on XMPP for realtime communication.

Hi and welcome to devnet, AFAIK, it's not on the radar yet. You can check the bug report system for that and if there's indeed nothing, you can open a new feature request about that topic.