Unsolved mapToGlobal, wrong values with Linux

[gmat]

So I created a simple stupid program that creates a main window and prints out the mapToGlobal values, the MainWindow snipit:

#include "mainwindow.h"
#include "ui_mainwindow.h"

MainWindow::MainWindow(QWidget *parent)
    : QMainWindow(parent)
    , ui(new Ui::MainWindow)
{
    ui->setupUi(this);
}

MainWindow::~MainWindow()
{
    delete ui;
}


void MainWindow::on_testPushButton_clicked()
{
    qDebug() << "MainWindow mapToGlobal" << this->mapToGlobal(QPoint(0,0));
}

On Windows and macOS this works perfectly. On a fresh install of Ubuntu 22.04 I move the window around the screen and get values like:

14:48:08: Debugging /home/user/projects/test/build/test ...
MainWindow mapToGlobal QPoint(12,120)
MainWindow mapToGlobal QPoint(180,1800)
MainWindow mapToGlobal QPoint(552,5520)
MainWindow mapToGlobal QPoint(654,6540)
MainWindow mapToGlobal QPoint(831,8310)
MainWindow mapToGlobal QPoint(900,9000)
MainWindow mapToGlobal QPoint(1002,10020)

Just so you know, the display resolution is 2048x1280.

Has anyone else experienced this ?

[mchinand]

What is the value of this->pos() at the same time you call mapToGlobal(QPoint(0,0))? I woud think it should be close but will differ slightly since I think pos() includes the Window frame when reporting the position when the widget is a window.

[gmat]

yes, it is not much different

    qDebug() << "MainWindow mapToGlobal" << this->mapToGlobal(QPoint(0,0));
    qDebug() << "MainWindow Pos" << this->pos();

output:

MainWindow mapToGlobal QPoint(6,60)
MainWindow Pos QPoint(3,30)

BTW I am using Qt 6.3.1

[JonB]

@mchinand
The point is that @gmat is showing values where in every case the Y coordinate == 10 x the X coordinate, which is pretty noticeable and odd... !

[mchinand]

Right, I'd like to see what pos() shows in those cases when the mapToGlobal() is showing those crazy y values. His single example above shows reasonable values for both mapToGlobal() and pos() (depending on where the window actual is, I guess).

[gmat]

Ah, so I went back to Qt 5.15.2 and it works as expected. Looks like they "introduced" a bug in 6.3.1.

[Asperamanca]

This one?

[gmat]

@Asperamanca said in mapToGlobal, wrong values with Linux:

Yup that looks like it.