Initializing a base class

class base
{
public:
base(const base &other) { // Init. here... }
// Stuff
};

class derived: public base
{
public:
derived(const derived &other): base(???) {}
};


What is the proper syntax for the initializer list in the derived
constructor to initialize its base subobject from the base subobject of
other?

re: Initializing a base class

Dave wrote:[color=blue]
> class base
> {
> public:
> base(const base &other) { // Init. here... }
> // Stuff
> };
>
> class derived: public base
> {
> public:
> derived(const derived &other): base(???) {}
> };
>
>
> What is the proper syntax for the initializer list in the derived
> constructor to initialize its base subobject from the base subobject of
> other?[/color]

See TC++PL/p306. Call a constructor of the base class, as though the
base class's name were the name of a member variable.

class Base
{
public:
Base( int );
};

class Derived: Base
{
public:
Derived( ):
Base( 3 )
{

}
};

re: Initializing a base class

"Jeff Schwab" <jeffplus@comcast.net> wrote in message
news:sM2dncXTsZ6Gh3eiRVn-tA@comcast.com...[color=blue]
> Dave wrote:[color=green]
> > class base
> > {
> > public:
> > base(const base &other) { // Init. here... }
> > // Stuff
> > };
> >
> > class derived: public base
> > {
> > public:
> > derived(const derived &other): base(???) {}
> > };
> >
> >
> > What is the proper syntax for the initializer list in the derived
> > constructor to initialize its base subobject from the base subobject of
> > other?[/color]
>
> See TC++PL/p306. Call a constructor of the base class, as though the
> base class's name were the name of a member variable.
>
> class Base
> {
> public:
> Base( int );
> };
>
> class Derived: Base
> {
> public:
> Derived( ):
> Base( 3 )
> {
>
> }
> };
>[/color]

But I'm talking specifically about the case of a copy constructor. Passing
an int is easy, but I don't know how to pass the base subobject of parameter
"other". Basically, I need to fill in the "???" as it appears in my
original post.

re: Initializing a base class

Dave wrote:[color=blue]
> "Jeff Schwab" <jeffplus@comcast.net> wrote in message
> news:sM2dncXTsZ6Gh3eiRVn-tA@comcast.com...
>[color=green]
>>Dave wrote:
>>[color=darkred]
>>>class base
>>>{
>>> public:
>>> base(const base &other) { // Init. here... }
>>> // Stuff
>>>};
>>>
>>>class derived: public base
>>>{
>>> public:
>>> derived(const derived &other): base(???) {}
>>>};
>>>
>>>
>>>What is the proper syntax for the initializer list in the derived
>>>constructor to initialize its base subobject from the base subobject of
>>>other?[/color]
>>
>>See TC++PL/p306. Call a constructor of the base class, as though the
>>base class's name were the name of a member variable.[/color][/color]

<snip>code</>
[color=blue]
> But I'm talking specifically about the case of a copy constructor. Passing
> an int is easy, but I don't know how to pass the base subobject of parameter
> "other". Basically, I need to fill in the "???" as it appears in my
> original post.[/color]

How do you expect to construct an object that can be constructed only
from an existing object of the same type? Unless the "Stuff" you
mentioned includes another constructor, I think you've got a design flaw
here. Of course, if you really want to walk on the dark side, you could
do something like this:

struct B { B( B const& ) { } };
struct D: B { D( ): B( *this ) { } };


-Jeff

re: Initializing a base class

"Jeff Schwab" <jeffplus@comcast.net> wrote in message
news:orKdnRMSNegAgHeiRVn-jw@comcast.com...[color=blue]
> Dave wrote:[color=green]
> > "Jeff Schwab" <jeffplus@comcast.net> wrote in message
> > news:sM2dncXTsZ6Gh3eiRVn-tA@comcast.com...
> >[color=darkred]
> >>Dave wrote:
> >>
> >>>class base
> >>>{
> >>> public:
> >>> base(const base &other) { // Init. here... }
> >>> // Stuff
> >>>};
> >>>
> >>>class derived: public base
> >>>{
> >>> public:
> >>> derived(const derived &other): base(???) {}
> >>>};
> >>>
> >>>
> >>>What is the proper syntax for the initializer list in the derived
> >>>constructor to initialize its base subobject from the base subobject of
> >>>other?
> >>
> >>See TC++PL/p306. Call a constructor of the base class, as though the
> >>base class's name were the name of a member variable.[/color][/color]
>
> <snip>code</>
>[color=green]
> > But I'm talking specifically about the case of a copy constructor.[/color][/color]
Passing[color=blue][color=green]
> > an int is easy, but I don't know how to pass the base subobject of[/color][/color]
parameter[color=blue][color=green]
> > "other". Basically, I need to fill in the "???" as it appears in my
> > original post.[/color]
>
> How do you expect to construct an object that can be constructed only
> from an existing object of the same type? Unless the "Stuff" you[/color]

That's exactly what copy constructors are all about!!!
[color=blue]
> mentioned includes another constructor, I think you've got a design flaw
> here. Of course, if you really want to walk on the dark side, you could
> do something like this:
>
> struct B { B( B const& ) { } };
> struct D: B { D( ): B( *this ) { } };
>
>
> -Jeff
>[/color]

Surely there is a correct way to copy construct derived objects. I'm only
trying to find out how it should properly be done! Let me modify the code
snippet you offered above:

struct B { B( B const& ) { } };
struct D: B { D( D const &other): B( ??? ) { } };

If I wanted to fill in the "???" such that initialization of B occurrs from
the B subobject of "other", how would I do so? other::B doesn't work,
other.B doesn't work, etc...

re: Initializing a base class

Dave wrote:[color=blue]
> "Jeff Schwab" <jeffplus@comcast.net> wrote in message
> news:orKdnRMSNegAgHeiRVn-jw@comcast.com...
>[color=green]
>>Dave wrote:
>>[color=darkred]
>>>"Jeff Schwab" <jeffplus@comcast.net> wrote in message
>>>news:sM2dncXTsZ6Gh3eiRVn-tA@comcast.com...
>>>
>>>
>>>>Dave wrote:
>>>>
>>>>
>>>>>class base
>>>>>{
>>>>> public:
>>>>> base(const base &other) { // Init. here... }
>>>>> // Stuff
>>>>>};
>>>>>
>>>>>class derived: public base
>>>>>{
>>>>> public:
>>>>> derived(const derived &other): base(???) {}
>>>>>};
>>>>>
>>>>>
>>>>>What is the proper syntax for the initializer list in the derived
>>>>>constructor to initialize its base subobject from the base subobject of
>>>>>other?
>>>>
>>>>See TC++PL/p306. Call a constructor of the base class, as though the
>>>>base class's name were the name of a member variable.[/color]
>>
>><snip>code</>
>>[color=darkred]
>>>But I'm talking specifically about the case of a copy constructor.[/color][/color]
>
> Passing
>[color=green][color=darkred]
>>>an int is easy, but I don't know how to pass the base subobject of[/color][/color]
>
> parameter
>[color=green][color=darkred]
>>>"other". Basically, I need to fill in the "???" as it appears in my
>>>original post.[/color]
>>
>>How do you expect to construct an object that can be constructed only
>>from an existing object of the same type? Unless the "Stuff" you[/color]
>
>
> That's exactly what copy constructors are all about!!![/color]

To *copy* an object, you first have to *have* an object. How do you
create the first such object, if the class has only a copy-constructor?
[color=blue][color=green]
>>mentioned includes another constructor, I think you've got a design flaw
>>here. Of course, if you really want to walk on the dark side, you could
>>do something like this:
>>
>>struct B { B( B const& ) { } };
>>struct D: B { D( ): B( *this ) { } };
>>
>>
>>-Jeff
>>[/color]
>
>
> Surely there is a correct way to copy construct derived objects. I'm only
> trying to find out how it should properly be done! Let me modify the code
> snippet you offered above:
>
> struct B { B( B const& ) { } };
> struct D: B { D( D const &other): B( ??? ) { } };
>
> If I wanted to fill in the "???" such that initialization of B occurrs from
> the B subobject of "other", how would I do so? other::B doesn't work,
> other.B doesn't work, etc...[/color]

The D is a B, so just do this:

struct D: B { D( D const &other): B( other ) { } };

Hth,
Jeff

re: Initializing a base class

"Jeff Schwab" <jeffplus@comcast.net> wrote in message
news:sO2dnbxvVNlnsnei4p2dnA@comcast.com...[color=blue]
> Dave wrote:[color=green]
> > "Jeff Schwab" <jeffplus@comcast.net> wrote in message
> > news:orKdnRMSNegAgHeiRVn-jw@comcast.com...
> >[color=darkred]
> >>Dave wrote:
> >>
> >>>"Jeff Schwab" <jeffplus@comcast.net> wrote in message
> >>>news:sM2dncXTsZ6Gh3eiRVn-tA@comcast.com...
> >>>
> >>>
> >>>>Dave wrote:
> >>>>
> >>>>
> >>>>>class base
> >>>>>{
> >>>>> public:
> >>>>> base(const base &other) { // Init. here... }
> >>>>> // Stuff
> >>>>>};
> >>>>>
> >>>>>class derived: public base
> >>>>>{
> >>>>> public:
> >>>>> derived(const derived &other): base(???) {}
> >>>>>};
> >>>>>
> >>>>>
> >>>>>What is the proper syntax for the initializer list in the derived
> >>>>>constructor to initialize its base subobject from the base subobject[/color][/color][/color]
of[color=blue][color=green][color=darkred]
> >>>>>other?
> >>>>
> >>>>See TC++PL/p306. Call a constructor of the base class, as though the
> >>>>base class's name were the name of a member variable.
> >>
> >><snip>code</>
> >>
> >>>But I'm talking specifically about the case of a copy constructor.[/color]
> >
> > Passing
> >[color=darkred]
> >>>an int is easy, but I don't know how to pass the base subobject of[/color]
> >
> > parameter
> >[color=darkred]
> >>>"other". Basically, I need to fill in the "???" as it appears in my
> >>>original post.
> >>
> >>How do you expect to construct an object that can be constructed only
> >>from an existing object of the same type? Unless the "Stuff" you[/color]
> >
> >
> > That's exactly what copy constructors are all about!!![/color]
>
> To *copy* an object, you first have to *have* an object. How do you
> create the first such object, if the class has only a copy-constructor?
>[/color]

Implicit default constructor.
[color=blue][color=green][color=darkred]
> >>mentioned includes another constructor, I think you've got a design flaw
> >>here. Of course, if you really want to walk on the dark side, you could
> >>do something like this:
> >>
> >>struct B { B( B const& ) { } };
> >>struct D: B { D( ): B( *this ) { } };
> >>
> >>
> >>-Jeff
> >>[/color]
> >
> >
> > Surely there is a correct way to copy construct derived objects. I'm[/color][/color]
only[color=blue][color=green]
> > trying to find out how it should properly be done! Let me modify the[/color][/color]
code[color=blue][color=green]
> > snippet you offered above:
> >
> > struct B { B( B const& ) { } };
> > struct D: B { D( D const &other): B( ??? ) { } };
> >
> > If I wanted to fill in the "???" such that initialization of B occurrs[/color][/color]
from[color=blue][color=green]
> > the B subobject of "other", how would I do so? other::B doesn't work,
> > other.B doesn't work, etc...[/color]
>
> The D is a B, so just do this:
>
> struct D: B { D( D const &other): B( other ) { } };
>
> Hth,
> Jeff
>[/color]

re: Initializing a base class

[color=blue][color=green]
>>To *copy* an object, you first have to *have* an object. How do you
>>create the first such object, if the class has only a copy-constructor?
>>[/color]
>
>
> Implicit default constructor.[/color]

If you declare any constructor for your class, the default constructor
will not be generated. Specifically, if you declare only a copy
constructor, then there is no implicit no-arg constructor.

re: Initializing a base class

"Jeff Schwab" <jeffplus@comcast.net> wrote in message
news:y_2dnfDYeInr2HeiRVn-sA@comcast.com...[color=blue]
>[color=green][color=darkred]
> >>To *copy* an object, you first have to *have* an object. How do you
> >>create the first such object, if the class has only a copy-constructor?
> >>[/color]
> >
> >
> > Implicit default constructor.[/color]
>
> If you declare any constructor for your class, the default constructor
> will not be generated. Specifically, if you declare only a copy
> constructor, then there is no implicit no-arg constructor.
>[/color]

Oh, of course! My bad... So, just assume it was part of "stuff"...

In any case, my question has been answered - I should have just realized
that I could pass "other" directly to the base class constructor and that it
would simply be sliced... Thanks for clearing it all up!