Might be permission issues, happened to me. I deleted the directory where its suppose to create it and it worked for me. Let the app create the directory.

@Damian7546
Not sure what your point is now. I only know/think that you could use that Diffie-Hellman stuff or similar from something in Open SSL, or you can Google for standalone implementations of it to write in C++ or you could just improve the approach in your own code for determining primes along the lines I suggested.

@SimonSchroeder said in Continuation not working with QtConcurrent::mapped:

What we have figured out so far is that mapped() returns a future. The second mapped (if it somehow would work) would wait for the full future to complete.

Why? That would be up to the implementation. It could track individual results (if there are many), and start those that can start.

Also remember that thread communication/synchronization has heavy overhead. If you chain two different mapped calls together you will have this overhead twice. This means it is quite sensible to put the two function calls into a single mapped call. I can't think of anything in the STL to merge two function calls to be passed as a new function pointer. You could come up with something like this (untested!):

Overhead, sure. It's my responsibility to only start work this way where it makes sense to move it to separate threads. My example is certainly not a good candiate :-)

In general, I wouldn't mind if ".then" on a "mapped" future simply wouldn't compile (best), or the behavior was documented and I got a runtime error (also ok). That is silently drops everything but the first result is "unexpected behavior".

I wonder about the design decision to make a QFuture both a single value AND a container...but I guess, that ship sailed long ago.

@JonB said in How to prevent checkbox text from overlapping when multiple checkboxes in QHBoxLayout?:

rather larger font that yours

Then there is no reproducible example above :)
And you cannot assume from the description that this is an issue only on Mac :)

@JonB ok, I changed my class so now it is a subclass of QStyledItemDelegate but nothing has changed. The behavior is the same.

By the way, I tried to reimplement displayText, but since it does not provide an index (just the text to be printed) I cannot retrieve the associated data from the model.

Maybe using containmentMask can help.

Try the following:

contentItem: Column { containmentMask: label Text { id: label } Text { id: description } }

Probably either the modifications aren't being done correctly in the underlying tree model, or the arguments to the beginMoveRows/beginRemoveRows are not correct. I haven't had enough coffee yet to be sure whether the latter are correct but they look as if they are at least along the right lines.

I would start off - if you haven't done so already - by checking that the state of the underlying tree is exactly as expected after each change.

@Christian-Ehrlicher Thanks Christian!

You're welcome !

You know, inverting lines can happen quickly and innocuously :-)
Anyway, now you're back on track.

Hi,

Your analysis looks correct. The default QAudioOutput constructor does indeed use the default output but in this case, it's the QAudioDevice constructor that is called.

You should check the bug report system to see if there's already something about and if not, please open a new ticket.

I think I finally understand how .ui.qml, .qml and python backend should work together when working with QDS project.
Way I did this was to export component ID as property alias in QDS by clicking @ sign, as shown on screenshot:
Screenshot 2025-02-05 181658.png

Then this alias should be used in .qml files, such as App.qml in order to access the component in .ui.qml (This was the biggest confusion point for me). Then we can use signals and slots to connect components with backend as usual.

I created a minimal example and can be seen here in case anyone has same issue.
https://github.com/irakliskhirtladze/QML-Demo

Hi, it's a bit hard to tell what is causing the issue without more information. Could you describe how the files stored in ${QML_FILES} are structured? Do you have QML files which are part of your module and which explicitly import the module itself?

But generally, this sounds like a bug (in Qt itself) – if only to improve the warning message – and I'd be interested in a bug report.

Lastly, as a workaround to avoid warnings, you could also set the "NO_GENERATE_EXTRA_QMLDIRS" option instead of setting the policy to old (if the Qt version is high enough).

@Christian-Ehrlicher Error of qt5 has been resolved thanks.

The new error arises as QGIS Entry Point Not Found
Screenshot (12).png

@Ejok88 Hi
I used the the book of shaders site tutorials to create the fuzzy circle check here https://thebookofshaders.com/
I actually got a very simple idea for your problem, just use a line with width which is just nothing more that a rectangle with two colors on the edjes, and use GL_TRIANGLE_STRIP , see this new example https://bitbucket.org/joaodeusmorgado/opengltutorials/src/master/T11_FuzzyLine/mygl.cpp#lines-185
Then you just add vertices to compose your line, note that you have to use "two" lines side by side to make the double fade out effect. Then you need to add 3 more lines without width, to the center and the 2 lines in the extremes. A bit of work but doable. I can help further if you need, but right now I just to much work overloaded. What is your use case, are you using widgets and Opengl ? I think going with quick3d would be easier, it that is an option for you.

13天了,连个模棱两可的敷衍都没有,真是挺高傲的.

I can develop Android and IOS apps using Qt and I love Texas.
I can sent you my CV via email.

Thanks for the feedback and bug reports !