[Qt-interest] Weird compiler error with QString(QChar(value))

[Stephen Chu]

In article <hrul12$9l1$1 at eple.troll.no>,
 Nikos Chantziaras <realnc at arcor.de> wrote:

> On 05/06/2010 05:35 PM, Stephen Chu wrote:
> > In article<201005052127.23761.thiago at kde.org>,
> >   Thiago Macieira<thiago at kde.org>  wrote:
> >
> > Thanks for the explanation. I ran into the same problem just last night.
> > Pulling my hairs out on this one.
> >
> > One question though. If that line of code actually declares a function,
> > what does 'unicodeValue' in the declaration do? I thought a function
> > declaration is something like:
> >
> >      QString s(QChar unicodeValue);
> 
> It's the same since you can put parentheses pretty much everywhere.  As 
> I posted previously for example, you can define main() like this:
> 
>    int main( int(argc), char**(argv) )
> 
> and it's accepted by the compiler.  That means that these two function 
> declarations:
> 
>    QString s(QChar unicodeValue);
>    QString s(QChar(unicodeValue));
> 
> are equivalent.  IMHO that's wrong, but then again, I don't know what 
> would break in the C++ standard if it would treat it as wrong.  But 
> given that C++ allows you to use simple types (int, char, etc) as classes:
> 
>    int i = int(50); //'int' is not a class, but you can use it as one
> 
> it should bark at the above definition of main().

Got it. Thanks. I was pulling my hairs when something like:

QRectF   r(QRectF(aPixmap.rect()));

works but this doesn't:

QRectF   r(QRectF(aRect));

Learn another trick (or trap?). The compiler should at least put a more 
informative warning on that line.

-- 
Stephen Chu