[PyQt] Floats accepted for C++ ints?

[Phil Thompson]

On 5 Jul 2016, at 9:51 pm, Florian Bruhin <me at the-compiler.org> wrote:
> 
> * Phil Thompson <phil at riverbankcomputing.com> [2016-07-05 18:33:37 +0100]:
>> On 5 Jul 2016, at 6:22 pm, Florian Bruhin <me at the-compiler.org> wrote:
>>> 
>>> I've noticed Python floats are accepted when a C++ int is expected,
>>> and simply truncated:
>>> 
>>> 	>>> from PyQt5.QtCore import QRect
>>> 	>>> QRect(0.5, 0.5, 0.5, 0.5)
>>> 	PyQt5.QtCore.QRect()
>>> 	>>> QRect(1.5, 1.5, 1.5, 1.5)
>>> 	PyQt5.QtCore.QRect(1, 1, 1, 1)
>>> 
>>> Is this intended behaviour? I'd have expected to raise a TypeError
>>> (or, if anything, accept floats where .is_integer() is true)
>> 
>> It's standard Python behaviour. Floats implement __int__().
> 
> They also implement __str__, yet I obviously can't use them where Qt
> expects a QString ;)
> 
> Isn't __int__ just there to customize what should happen when using
> int(), just like __str__ is for str()?
> 
> float is not a subclass of int (i.e. isinstance(1.5, int) == False),
> which is why this not raising a TypeError confuses me.

It's how PyLong_AsLong() works...

https://docs.python.org/3/c-api/long.html#c.PyLong_AsLong

Phil