qml <-> c++ context communication options

[solemn shell]

any one with qml experiences?
i have idea how to communicate between qml and c++ context in way i like
what can you say?

class FrontendConnectionSide{
Q_OBJECT
Q_PROPERTY(SomeType notifiable READ getNotifiable WRITE setNotifiable NOTIFY notifyBackend)
public:
FrontendConnectionSide(BackendConnectionSide* back){
// connect notify to backend with queued con, without magic
// connect(this,&notifyBackend,backend,&receiveFunc,Qt::queued...)
assert(back);
back_con = back;
// binded qthread of backend may be different, but if use non direct for backendCon <-> backendModules it is negligible
}

public:
// prop set get...

public slots:
void sendSomethingToBackend(){
   // my macros
   INVOKE(back, receiveSomethingFromFrontend);
}
// etc...

signals:
//properties signals...

signals:
void receiveSomethingFromBackend();

private:
SomeType notifiable;
BackendConnectionSide* back_con;

};

////

int main(){
    BackendConnectionSide* back = ...;
    FrontendConnectionSide frontendConnectionSide(back);
    engine.rootContext()->setContextProperty("frontendConnectionSide", &frontendConnectionSide);
    // connect back slots/signals to logic modules, may be in different threads
}

///////////
Item {
    Text {
        text: frontendConnectionSide.notifiable // changes will trigger something
    }

    Button {
        onClicked: frontendConnectionSide.sendSomethingToBackend()
    }

    frontendConnectionSide.receiveSomethingFromBackend.connect(function {
         ///
    })
   // or proper style
/*
    Connections {
        target: applicationData
        onDataChanged: console.log("The application data changed!")
    }
*/
}```

[forest olive]

When your question is answered use !solved to mark the question as resolved.

Remember to ask specific questions, provide necessary details, and reduce your question to its simplest form. For tips on how to ask a good question use !howto ask.

[solemn shell]

i like this style of communication, it is literally web without extra overhead, and also thread independent at all, if get rid of extra queued invoke then without overhead at all
i use something similar for widget app
but it just object DeepConnections:singleton{signals: void fromObject_toObject(int);}
with qml im need bridge between js and c++...
.
btw i learned something now with chatbot, may contain principal issues
struggle with normal qml tutorials

there is more options, but if possible i would like to use only frontend-backend interface
https://doc.qt.io/qt-6/qtqml-cppintegration-overview.html

[solemn shell]

!solved

[forest olive]

Thank you and let us know if you have any more questions!

This thread is now set to auto-hide after an hour of inactivity

[vital rampart]

Looks interesting, but doesnt this go to #1078717238678409369

[solemn shell]

no because it is conceptual question
also solved on qt server, they said i should use qml singlton because it right api for that, but even with context its fine

[bitter aurora]

assert is not a function, it's a macro, open the raw macro definition and see why it's just bad, don't use it

[solemn shell]

wtf are you talking about?

[bitter aurora]

/* void assert (int expression);

   If NDEBUG is defined, do nothing.
   If not, and EXPRESSION is zero, print an error message and abort.  */

#ifdef    NDEBUG

# define assert(expr)        (__ASSERT_VOID_CAST (0))

/* void assert_perror (int errnum);

   If NDEBUG is defined, do nothing.  If not, and ERRNUM is not zero, print an
   error message with the error text for ERRNUM and abort.
   (This is a GNU extension.) */

# ifdef    __USE_GNU
#  define assert_perror(errnum)    (__ASSERT_VOID_CAST (0))
# endif

[solemn shell]

and what??

are you trolling or something?

i have like 500+ asserts in my codebase

[bitter aurora]

in release builds, your assert is a no-op

[solemn shell]

ok, ok i got it, you just troll

[bitter aurora]

#ifdef    NDEBUG

# define assert(expr)        (__ASSERT_VOID_CAST (0))

it says it here

and you are still rambling

for no reason

[steep barn]

assert gets deleted in release builds LITERALLY
So no no-op gets done. since its processed and literally removed

[bitter aurora]

well, a no-op I don't mean the literall assembly instruction for the no-op operation, mainly used for padding stuff, I mean that, it just gets deleted, yes

[steep barn]

so... it doesn't really matter for release builds and works for debug to debug stuff which is very important

[bitter aurora]

oh and when using some, custom made macros for, checking an invariant, the macro may be something not from <assert.h> but some custom made header, with a custom made macro, that does something else, so at the end of the day, check if the thing is a macro or function, and then reveal its source to find out what it does

we have a custom made assertions library at our work, and it's very good, I think

[steep barn]

yes

cool