General and Desktop

The issue was auto solved because there has been no feedback to the request for tests. If you can do what was suggested there and add some more information to it then the bug can be reopen.

@Kent-Dorfman It is solved.

mark as solved?

YMMV inserting images into a database. BLOB field support is often a special case, if supported by the DB at all.

@JonB @JoeCFD Okay I just switched to using QMessageBox then, I had not thought about running this on different operating systems. Thank you for the tip. @Christian-Ehrlicher thank you for the textChanged input. I am going to try to implement that

@Alby Download binary and headers. In your .pro file add INCLUDEPATH, DEPENPATH and LIBS manually or as described in the manual. After that you can #include headers in your sources. If you do #include but QtCreator complains that file can't be found - you did the above wrong.

@CJha Yes, exactly that.

@poordev123 Just so you know: it's QObject::findChildren() so you can actually call it on any QObject, and that includes any and all QWidgets. QGroupBox is a QWidget itself, and there are other such "container" widgets for which this works.

@Habibie said in Qt dynamic translation on a serialized enum key in a comboBox: If so, I noticed, the lupdate utility does not pickup all the enum keys and I reckon that's the problem. Because you do not properly define them - see QT_TRANSLATE_NOOP documentation.

@Christian-Ehrlicher Thank you. I've seen that and double checked that MYSQL_VERSION_ID is correct. It appears as 80026 on the latest MySQL 8.0.26.

@J-Hilk said in How to pass signal from one class to another.: you can connect without an receiver QObject You can indeed, but you cannot use that to connect to instances of the OP's Geometry, which is what he is trying to do. He might write, e.g., a static method in Geometry to use as the slot, but it won't then do what he has in mind, which is to "auto-connect" each instance created to main window's signal.

@jsulm Thanks.i can finally figure it out .i thought my question is too stupid when i asked this in stackoverflow.but it is ok now.Thanks for your help!

@Publicnamer It depends just what you want, what is in your model, and whether you know how much is in the backing model. Your question seems to relate to paging the data, or something based off that. Rather than looking at visual rects --- especially if they don't work --- the alternative is to look at the scrollbar. That reflects how far through the data the user is. One possible implementation is given in Pagination in QtableWidget, @raven-worx's algorithmic outline there. I have seen others (unfortunately can't find references) where the algorithm is to maintain the scrollbar height such that there is always a "gap" at the bottom which the user can scroll into, and that fires a fetch of more data and recalibration of the scrollbar, which still has a "gap" at the bottom. Think of when you use an application which produces large or endless output, like a web page receiving more and more data or showing photos where more and more are to come (e.g. some tumblr pages). The scrollbar always allows the user to scroll some more. @raven-worx's suggestion could be adapted to that. If you Google for stuff involving QTableWidget or QTableView and pagination you may find some useful code/ideas.

@IknowQT You can define build folders for debug and release in QtCreator under Projects/Build & Run. Also, I would remove CONFIG += debug, debug/release mode is handled in QtCreator.

@Publicnamer Yes. I only point out the limitations of your solution for future readers. I am not saying that it is bad for you but I only warn the possible inconveniences for other users.

@mattfbacon I find that QT_SCALE_FACTOR=2 does an amazing job at scaling. But I'm also wondering, is it real scaling at the pixel level, or is it just an OpenGL trick that makes that window bigger but the number of pixels in the window is unchanged?

@jazzco2 Right, but what's the solution? If an older windows dll fixes things, I could try finding and swapping it P.S. that thread talks about a different issue. It is a crash not a warning with strange behaviour and there is no solution there either

@Bonnie The real problem here is that a screen of any size can be used like a desktop, or a phone screen. For instance the Raspberry Pi 7" is always used like a desktop UI. Thus it is a question of "user interface idiom" i.e. how the screen is meant to be used.

Yes and no - it seems to have been solved in an update. I checked it right now and the issue does not persist. I did not really find out what caused the problems.