[PyQt] QPoint not automatically cast to QPointF

Shriramana Sharma samjnaa at gmail.com
Thu Sep 20 12:19:46 BST 2012


Hi. Thanks for your reply. What you have to do in SIP, is it of the nature
of a workaround to Python's inability to do the conversion automatically?
Or Python is able to do the conversion automatically and SIP just needs to
present the classes in a proper manner as required by Python?

Sent from my Android phone
On Sep 20, 2012 2:19 PM, "Phil Thompson" <phil at riverbankcomputing.com>
wrote:

> On Thu, 20 Sep 2012 07:34:10 +0530, Shriramana Sharma <samjnaa at gmail.com>
> wrote:
> > In my recent work with Beziers I ran across this. A QPoint is *not*
> > automatically converted into a QPointF in Python/PyQt while it *is*
> > converted in C++:
> >
> > The following C++ code compiles fine:
> >
> > # include <QtCore/QPoint>
> > # include <QtGui/QPainterPath>
> > int main ( void ) {
> >       QPainterPath p ;
> >       QPoint p1 ( 100, 150 ), c1 ( 166, 250 ), c2 ( 234, 250 ), p2 (
> 300, 150
> )
> >       ;
> >       p . moveTo ( p1 ) ;
> >       p . cubicTo ( c1, c2, p2 ) ;
> > }
> >
> > whereas its Python/PyQt equivalent:
> >
> > #! /usr/bin/env python3
> > from PyQt4 . QtCore import QPoint
> > from PyQt4 . QtGui import QPainterPath
> > p = QPainterPath ()
> > p1 = QPoint ( 100, 150 )
> > c1 = QPoint ( 166, 250 )
> > c2 = QPoint ( 234, 250 )
> > p2 = QPoint ( 300, 150 )
> > p . moveTo ( p1 )
> > p . cubicTo ( c1, c2, p2 )
> >
> > produces the following:
> >
> > Traceback (most recent call last):
> >   File "./qpoint-test.py", line 11, in <module>
> >     p . moveTo ( p1 )
> > TypeError: arguments did not match any overloaded call:
> >   QPainterPath.moveTo(QPointF): argument 1 has unexpected type 'QPoint'
> >   QPainterPath.moveTo(float, float): argument 1 has unexpected type
> >   'QPoint'
> >
> > But QPointF in PyQt *does* provide a constructor from QPoint, so is
> > this the limitation of Python that it does not automatically check
> > whether it can convert one type to another to satisfy a function's
> > call signature? (Or is there some other fault in my PyQt code?)
> >
> > I realize I could always convert it manually using QPointF(p1) etc but
> > the assumption seems natural that an integer-precision numeric object
> > should be accepted where a float-precision can...
>
> It requires a change to SIP to handle this automatically which is on the
> TODO list but I haven't got round to it yet.
>
> Phil
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.riverbankcomputing.com/pipermail/pyqt/attachments/20120920/32d77f7b/attachment.html>


More information about the PyQt mailing list