General and Desktop

Still cannot post code "Post content was flagged as spam by Akismet.com" The code in the image gives me --- 1.1: map true isMapped true isWritable true --- 1.2: map true isMapped false isWritable false --- 2.0: map true isMapped true isWritable false --- 2.1: map false isMapped true isWritable false --- 2:2 map false isMapped false isWritable false Without any sleep() it does indeed work, but I mean in reality there is some delay between frames in a video, right? I feel there is a bug somewhere. [image: ee58c7dc-25af-4b1b-a392-9ae33ccea846.png]

@JonB Nothing to do with QSettings. It was a Unix locale issue. It turns out that the QSettings stuff was left there by the original developer of that code to allow testing without switching the language of his system. It mightily confused me ...

Hi @Zeshan, at first you should clarify if you want to use Modbus TCP or MQTT, as these are two completely different protocols. Regards

Here the examples of how to use CPP in Swift with XCode projects.

--bump--

@Christian-Ehrlicher thanks for replying. I've seen the document, but i don't understand why it's rendered like that. if i need a big rect contains a small image, did i should just do like this: QRect big(0,0,152,152); QRect small(1,1,150,150); painter.setPen(QPen(Qt::red,1)); painter.drawRect(0,0,big.width-1,big.height-1); painter.drawImage(1,1,small.width-1,small.height-1); // or // painter.drawImage(1,1,small.width,small.height); ??? width and height for drawRect and drawImage is not the same meaning? right?

I tried all sorts of things as well, searched all over, and eventually found this post https://stackoverflow.com/questions/68303194/why-adjustsize-doesnt-resize-mainwindow-in-qt In the end all that I needed to do was ui.centralWidget->adjustSize(); adjustSize(); When a widget was hidden, the window would resize to remove the unused space. When the widget was unhidden, the window and layout would resize to make space for the widget. This was done with Qt 6.5.5, but I would guess this would work with earlier versions as well.

Thank you so much ,it works that way ,on the other hand ,i dont limit the values because it is needed that way

@nicker-player I know nothing about QML or QResource. But since you are using bool QResource::registerResource(const uchar *rccData, const QString &mapRoot = QString()) overload it tells you Warning: The data must remain valid throughout the life of any QFile that may reference the resource data. That means you cannot afford to delete the newed array at all, so long as it may be referenced, does it not? You would need to save the newed pointer and only delete it after you no longer need the resource. I only say this from the description of that overload. If you could use the bool QResource::registerResource(const QString &rccFileName, const QString &mapRoot = QString()) overload instead you would not have this issue --- presumably because that one keeps the file open and gets the data from there instead of requiring the data to be sitting in permanent memory?

There are items in this main window that I don't need and can't use anyway, and I found those components in a script named "CameraSettingsPage dot py." I was able to delete the "Save RAW and Meta data" item easily, so I saved a whole line of space there, but when I deleted code for the "Sensor mode" (which also seems to be unusable)... # センサモードのコンボボックス self.sensorFormat = QComboBox() #self.sensorFormat.addItem("Default") self.sensorFormat.addItems([f'{x["formatfsensor"].format} {x["size"]}' for x in picam2s[camid].sensor_modes]) # 最大の解像度(4608x2592)に設定 # プレビューとキャプチャの画角違い防止のため固定 self.sensorFormat.setCurrentIndex(2) self.sensorFormat.setEnabled(False) formLayout.addRow("センサーモード", self.sensorFormat) ...the program would no longer launch. I assume there are references to things in this code elsewhere in this script (or maybe other scripts) that I would also need to delete or modify. I also would like to get rid of the "Image Size" (画像サイズ) item just below the sensor mode item. # 画像サイズ反映ボタン self.resButton = QPushButton("反映") self.resButton.setEnabled(False) # 画像サイズ反映ボタンクリック時の動作 self.resButton.clicked.connect(self.on_resButton_clicked) # 画像の縦横サイズを変更時は反映ボタンを有効化 self.imageWidth.valueChanged.connect(lambda: self.resButton.setEnabled(True)) self.imageHeight.valueChanged.connect(lambda: self.resButton.setEnabled(True)) # 画像サイズ resolution = QWidget() resHLayout = QHBoxLayout() resolution.setLayout(resHLayout) resHLayout.addWidget(self.imageWidth) resHLayout.addWidget(QLabel("x"), alignment=Qt.AlignHCenter) resHLayout.addWidget(self.imageHeight) resHLayout.addWidget(QLabel(" "), alignment=Qt.AlignHCenter) resHLayout.addWidget(self.resButton) formLayout.addRow("画像サイズ", resolution) Getting rid of those two would make this window plenty usable and legible. (As it is, the "Rotation" control, just below the "Image size" control, is unusable with the window sized to fit my screen, because the button isn't visible.)

@DungeonLords said in killTimer: Timers cannot be stopped from another thread: void Worker::run() processEvents() in an infinite loop is not the best solution here. This will produce code that will use a single core a 100%. People will start asking questions why your app is using 100% of one core while it is doing nothing (I'm speaking from experience). The proper solution is to call exec() instead. However, this is also already the default implementation of QThread::run(). So, you can just leverage that and throw away your entire while-loop. Maybe then don't call your method run() but just createTimer(). Then it makes a lot more sense to call connect(thread.get(), &QThread::started, worker.get(), &Worker::createTimer());, i.e. createTimer when thread is started. (Instead of using requestInterruption you would then just call quit.) The use of exit_LABEL shows some lack of understanding of asynchronous programming, which is the natural way in Qt. Send signals connected to slots instead to perform such operations. Just out of curiosity: Was AI heavily involved in creating this code? Or are you just coming from embedded systems programming? emit GetTimeAsk(); does not make sense. emit is usually used for signals only (it compiles because it is replaced by nothing by the preprocessor). Effectively you are just executing a normal member function. This also means that this is just executed in the main thread. (Which in turn means that incrementing i is not thread safe.)

@Pl45m4 said in Problem with deleting QBuffer: Simple: Link you release build to Qt release libs and your app's debug build to Qt's debug libs. Yes. Apparently I was linking external release libraries instead of debug ones. The issue is solved. Thank you!

@Scherer-Tech said in Use QStandardItemModel with QTreeView: how to connect the model to a QTreeView See https://doc.qt.io/qt-6/qabstractitemview.html#setModel and https://doc.qt.io/qt-6/model-view-programming.html

Final Problem Solution： https://blog.csdn.net/qq_32854345/article/details/128022276

@IgKh Thank you very much. That makes a lot of sence. I agree that the previous behavior can be considered a defect. I wonder why I didn't notice that before, but I think I was happy that it looked the same on all 3 platforms

@Pl45m4 I understand your point but how hubstaff , timedoctor,tahometer these apps track user keyboard mouse and even screens activity on other apps

@papisant Well, I'm not sure why it did not work with your original m_mediaDevices{std::make_shared<QMediaDevices>(this) which would surely be a class member too. Anyway, the important take away from this is: when connecting a signal you must look at the scope/lifetime of the signalling object (same for the slot). If either of these is a "local variable" that means the object will go out of scope and be destroyed at the end of the function they are in. And when Qt sees either the signalling or slot object being destroyed it auto-disconnects any connected signals/slots for that object (for obvious reasons). A connection only persists as long as both signal & slot objects are in existence. That is why class members are often used, or the object must be newed and not deleted.