https://bugs.kde.org/show_bug.cgi?id=358689

--- Comment #9 from Weng Xuetian <wen...@gmail.com> ---
(In reply to Martin Gräßlin from comment #8)
> (In reply to Weng Xuetian from comment #7)
> > This setting has nothing to do with X11, it's just for Qt itself.
> > 
> > you can set it like:
> > QT_SCREEN_SCALE_FACTORS=eDP1=2;eDP2=3
> > 
> > So when you move window among screen, the Qt's window will be rescaled to
> > the screen dpi automatically.
> 
> That cannot work on X11, sorry. That's just Qt's "I don't know how to
> multi-screen on X11". The problems here are:
> * a window is not connected with an output, it has a position within the X
> screen space
> * QWindow allows to set the QScreen, but this does not influence whether the
> window is sown on the screen
> * DPI for multi-screen on common X-Servers is broken. As soon as there are
> two screens, the X-Server does report a default 96 pdi AFAIK
> * when moving a window from one screen to another it could result in an
> infinite jumping between the different DPI modes
> * the totally legit situation of one window on two screens is not handled
> * the window manager does not adjust the window decoration
> * I could go on and on
> 
> Exposing a config option for that would be wrong: we cannot have different
> scale factors on X11. That would just introduce another set of things which
> can go wrong with Qt's utterly broken (and broken beyond repair)
> multi-screen implementation on X11.

Such complain doesn't solve anything. You're just pointing to some "Potential"
Problem that doesn't really happen. 

I have been using QT_DEVICE_PIXEL_RATIO=auto (when it's still old style env
var) since Qt 5.4, and it never cause me any trouble. Dynamically change
window's scale factor is tricky, but it's also the same approach on Windows. 

I have been use it for such a long time. And if any issue you claim actually
happens, Qt should fix it. Not to mention IT IS not the case. I have even seen
user complain about startkde script prevent user to set
QT_AUTO_SCREEN_SCALE_FACTOR to 1. It's not something badly broken, really.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to