[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-devel
Subject:    Re: strings and QStrings
From:       Patrick Julien <freak () desource ! net>
Date:       2000-12-10 13:04:32
[Download RAW message or body]

On December 10, 2000 05:26 am, Christopher Willmot wrote:
> On Sun, Dec 10, 2000 at 09:29:55AM +0100, Bo Thorsen wrote:
> > Date: Sun, 10 Dec 2000 09:29:55 +0100 (MET)
> > From: Bo Thorsen <bo@sonofthor.dk>
> > To: kde-devel@master.kde.org
> > Subject: Re: strings and QStrings
> >
> > On Sun, 10 Dec 2000, David van Hoose wrote:
> > See this code: ...
>
> Ah, I thought constructors and assignment operators were semantically
> different.  

They are, but in the following cases...

	B b1 = a;
	B b2 ( a );

these are not assignment, they are both initializers.  ISO says the copy 
constructors get called for init, and assignment operator get called for 
assignment.

So, after construction, ...
	b2 = a; 

the assignment operator gets called.

>If that's the case it would surely be wrong for the compiler
> to substitute one for the other.  Scott Meyers, in his (now rather old?)
> book "Effective C++" claims that the compiler has to create a temporary
> object which is then passed to the assignment operator.  That would
> account for there being two contructor calls.

Not in the code that was provided,  temporaries are made when conversions for 
parameters are needed.  Since B has a copy constructor that takes an A&, no 
other constructor is called.

>
> But aren't QStrings shared?  So either way is should be just a matter of
> increasing a count in the shared object.

Not in this case, it's the char * inside a string converted to const char * 
being assigned to a QString.
 
>> Visit http://master.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<

[prev in list] [next in list] [prev in thread] [next in thread]