[Development] Should QVariant be doing fuzzy comparisons on doubles?

[Mathias Hasselmann]

Am 21.09.2016 um 12:00 schrieb Marc Mutz:
> On Wednesday 21 September 2016 11:42:41 Mathias Hasselmann wrote:
>> No matter what order I use for config and value, the compiler will pad
>> and -Wpadded will complain. How am I supposed to fix this? This
>> solutions that come to my mind all are ugly, but most likely I am just
>> stupid.
>
> Tabs? Seriously? :)

E-mail clients and their horrible text editors.

>
>  	struct Sensor
>  	{
>  	    enum Config { NotInitialized, PullUp, PullDown, Analog };
>
>  	    Config config;
> 	    char padding[alignof(QVariant) - sizeof(Config)];
>  	    QVariant value; //[1]
>  	}
>
> and pray for alignof(QVariant) - sizeof(Config) != 0.

Yup, one of solutions I had in mind and rejected for expected that 
prayer remaining unheard while juggling with 32 and 64 bit code.

> I wasn't serious about -Wpadded, but incidentally, my example shows that type
> traits also won't help. We have a handful of classes in Qt which have a
> 'reserved' field that doesn't get initialized (so it better should be named
> 'unusable'). A type trait cannot tell if a field called 'padding', 'reserved',
> 'unused' or 'unusable' is meant to be part of equality or not.

Q_DECLARE_TYPEINFO(QBar, Q_PADDED_TYPE) // :D

Ciao,
Mathias