Qt 6

@Spazz have you solved the problem? It seems tha I met the same trouble that the vertexCount is alway 0.

@fawad
Like @jsulm says, I don't use your environment but if I did I would hope to do everything inside and only use one of either Creator or Visual Studio, swapping between them sounds like a pain!

The fact that it's a 32 bit target is not an issue. Depending on the library, the harder is often to make its build system use the cross-compiler on the host (things have vastly improved on that side since I was doing embedded work).

See Examples, Tests, and Tools
and Third-Party Libraries

I'm running into this in windows. I've been looking at the qmake -d 2>&1 of a Qt example vs a QWT example. At some point in the QWT qmake generation it seems to loose the entrypoint_private module.

qt example: DEBUG 1: C:/tools/Qt/6.4.2/msvc2019_64/mkspecs/features/qt_functions.prf:379: modules := core gui entrypoint_private
qwt example: DEBUG 1: C:/tools/Qt/6.4.2/msvc2019_64/mkspecs/features/qt_functions.prf:379: modules := core gui printsupport concurrent opengl openglwidgets svg

I haven't been able to figure out the clean fix yet. I ended up diffing the makefiles between the examples and one the major differences was that the QWT example was missing the lib: Qt6EntryPoint.lib
So I added it to the qwt example pro like:
LIBS += C:\Qt\6.4.2\msvc2019_64\lib\Qt6EntryPoint.lib

and then the release version at least built successfully.

note: for diffing the qmake -d output it's handy to do a cut like this (powershell):
qmake -d 2>&1 | % { $_ -split ':' | select -last 1 }

this strips out a lot of paths making the diff easier to scan.
This was using Qwt 6.2.0, git sha:92baef.

@IceMichael said in Precompiled headers with CMake in Qt 6. How?:

Interesting would be to see how we can a wellformed list of Qt header names that can be used to pass to the CMake function

No since CMake don't know what headers you're using / you want to use. Even when you link against Qt::Widgets adding the <QtWidgets> header to the precompiled target may be wrong. Adding some important headers manually shouldn't be that big problem.

@sierdzio Thank you a lot for your answer !

Your explaination on the opaque pointer is clear, I didn't have that notion before.

With this knowledge, most of my questions are wrong, like the 5th.

@Asperamanca confirmed. Thank you so much!

I think the best would be to take a look at the sources if you want to fully understand how it works.

@Chris-Kawa

Thank you for your answers. I followed your recommendations and everything is ok now.
Here is the corrected code

out << static_cast< float >(samples[i]) / std::numeric_limits<qint16>::max() << "\r\n";

@Dave-Hughes Es you can see from the error message your Python version is not supported yet. Use Python 3.11.

Addition: probably, this fault appears on 100% CPU load.