Forward References

[Randy Yates]

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}
results in
g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I /devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27
Why? I've declared the forward reference at the beginning of the file. Is
this just a stupidity of g++??
--
% Randy Yates % "Watching all the days go by...
%% Fuquay-Varina, NC % Who are you and who am I?"
%%% 919-577-9882 % 'Mission (A World Record)',
%%%% <ya***@ieee.org> % *A New World Record*, ELO
http://home.earthlink.net/~yatescr

[Randy Yates]

Randy Yates <ya***@ieee.org> writes:

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}
results in
g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I /devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27
Why? I've declared the forward reference at the beginning of the file. Is
this just a stupidity of g++??

I forgot to add that if I switch the order of the class declarations
the error goes away. I shouldn't have to order class declarations
which have forward references in my files, just as I don't have to
order subroutines if they have prototype definitions.
--
% Randy Yates % "The dreamer, the unwoken fool -
%% Fuquay-Varina, NC % in dreams, no pain will kiss the brow..."
%%% 919-577-9882 %
%%%% <ya***@ieee.org> % 'Eldorado Overture', *Eldorado*, ELO
http://home.earthlink.net/~yatescr

[JKops]

Randy Yates posted:

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}

Rearrange it:

(Plus I corrected a typo, it should've been "argv[]")
class DATE
{
public:
int day;
int month;
int year;
};

class MYCLASS
{
public:
int n;
DATE date;
};
int main(int argc, char* argv[])
{
return 0;
}
-JKop

[Julián Albo]

Randy Yates wrote:

Why? I've declared the forward reference at the beginning of the file. Is
this just a stupidity of g++??

Forward declaration allow only to declare pointer or references, to declare
an object of the class the complete declaration is required.

--
Salu2

[John Harrison]

"Randy Yates" <ya***@ieee.org> wrote in message
news:u0**********@ieee.org...

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}
results in
g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I
/devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27
Why? I've declared the forward reference at the beginning of the f ile.Is
this just a stupidity of g++??

C++ is no different from C in this regard. Replace class with struct, remove
'public:', and you'd have a C program and exactly the same error. You are a
C programmer, no?

In order to compile MYCLASS the compiler needs some basic information, such
as the size of each data member. It cannot get that from a forward
declaration.

class DATE;

class MYCLASS
{
public:
int n;
DATE* date;
};

This is OK because DATE* has a known size; but what you wrote is not because
the compiler does not know the size of DATE.

Suggest you simply reorder the classes.

john

[David Lindauer]

Randy Yates wrote:

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}

results in

g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I /devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27

Why? I've declared the forward reference at the beginning of the file. Is
this just a stupidity of g++??
--
% Randy Yates % "Watching all the days go by...
%% Fuquay-Varina, NC % Who are you and who am I?"
%%% 919-577-9882 % 'Mission (A World Record)',
%%%% <ya***@ieee.org> % *A New World Record*, ELO
http://home.earthlink.net/~yatescr

when the compiler is compiling 'myclass' it runs into a forward declared reference to 'date'. However, while date is known to
exist, the compiler does not know anything about it, such as size or layout. But at this phase of the compilation process, the
compiler needs to know at *least* the size, and possibly more so that it can complete the declaration of myclass. That means that
to use date here, you had to have previously done more than forward declaring it.

The forward declarations are primarily used to declare types that are going to be used as pointer references
in subsequent classes, this way you can have classes that point to each other in a circular fashion.

David

[Randy Yates]

"John Harrison" <jo*************@hotmail.com> writes:

"Randy Yates" <ya***@ieee.org> wrote in message
news:u0**********@ieee.org...

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}
results in
g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I
/devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27
Why? I've declared the forward reference at the beginning of the f ile.Is
this just a stupidity of g++??

C++ is no different from C in this regard. Replace class with struct, remove
'public:', and you'd have a C program and exactly the same error. You are a
C programmer, no?

Yes, and structure definitions have no such thing as a "forward reference," and
if they did I'd be asking the same question.

If a forward reference isn't good for this, then what is it good for?
--
% Randy Yates % "...the answer lies within your soul
%% Fuquay-Varina, NC % 'cause no one knows which side
%%% 919-577-9882 % the coin will fall."
%%%% <ya***@ieee.org> % 'Big Wheels', *Out of the Blue*, ELO
http://home.earthlink.net/~yatescr

[Mike Wahler]

"Randy Yates" <ya***@ieee.org> wrote in message
news:3c**********@ieee.org...

"John Harrison" <jo*************@hotmail.com> writes:

"Randy Yates" <ya***@ieee.org> wrote in message
news:u0**********@ieee.org...

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}
results in
g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I
/devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27
Why? I've declared the forward reference at the beginning of the f ile.Is this just a stupidity of g++??
C++ is no different from C in this regard. Replace class with struct, remove 'public:', and you'd have a C program and exactly the same error. You are a C programmer, no?

Yes, and structure definitions have no such thing as a "forward

reference,"

Nor do class definitions. But in C, yes a forward reference to
a struct can indeed be declared.

The following is valid C (as well as C++):

struct x; /* forward declare type 'struct x' */
struct x *p; /* define a pointer to type 'struct x' */

/* (but an actual object of type 'struct x' cannot be
created until after type 'struct x' is completely defined) */

struct x s; /* invalid */

/* fully define type 'struct x' */
struct x
{
int a;
};

struct x s2; /* valid */
and
if they did I'd be asking the same question.

If a forward reference isn't good for this, then what is it good for?

For defining a pointer (or reference in C++) to a struct or class which
has yet to be defined.

This issue works exactly the same way in C as in C++

-Mike

[John Harrison]

>>

C++ is no different from C in this regard. Replace class with struct,
remove
'public:', and you'd have a C program and exactly the same error. You are
a
C programmer, no?
Yes, and structure definitions have no such thing as a "forward
reference," and
if they did I'd be asking the same question.

Not true.

struct Node;

That is perfectly good C and C++.
If a forward reference isn't good for this, then what is it good for?

It tells the compiler that a name is the name of a class (or struct). You
can do a few things with just a name, for instance you can declare a pointer
or a reference. For instance

struct Node;

class List
{
public:
void add_node(Node*);
void remove_node(Node*);
};

John

[Randy Yates]

Thanks David, John, and Mike. I guess I see the point now, but
I don't see why the compiler couldn't have been designed to
forward scan the file and pick up that information without
the user having to go through these hoops - something like
a "two-pass" compiler design.

--RY
David Lindauer <ca*****@bluegrass.net> writes:

Randy Yates wrote:

I'm having a problem with forward references.

For example,

class DATE;
class MYCLASS;

class MYCLASS
{
public:
int n;
DATE date;
};

class DATE
{
public:
int day;
int month;
int year;
};

int main(int argc, char* argv)
{
return 0;
}

results in

g++ -c -DTARGET_PC -Wall -g -I /. -I /mingw/include/ -I /devstudio/vc/include -o ./pc/impmain.o -Wno-deprecated impmain.cpp
impmain.cpp:8: field `date' has incomplete type
make: *** [pc/impmain.o] Error 1

Compilation exited abnormally with code 2 at Sat Oct 16 12:54:27

Why? I've declared the forward reference at the beginning of the file. Is
this just a stupidity of g++??
--
% Randy Yates % "Watching all the days go by...
%% Fuquay-Varina, NC % Who are you and who am I?"
%%% 919-577-9882 % 'Mission (A World Record)',
%%%% <ya***@ieee.org> % *A New World Record*, ELO
http://home.earthlink.net/~yatescr

when the compiler is compiling 'myclass' it runs into a forward declared reference to 'date'. However, while date is known to
exist, the compiler does not know anything about it, such as size or layout. But at this phase of the compilation process, the
compiler needs to know at *least* the size, and possibly more so that it can complete the declaration of myclass. That means that
to use date here, you had to have previously done more than forward declaring it.

The forward declarations are primarily used to declare types that are going to be used as pointer references
in subsequent classes, this way you can have classes that point to each other in a circular fashion.

David

--
% Randy Yates % "My Shangri-la has gone away, fading like
%% Fuquay-Varina, NC % the Beatles on 'Hey Jude'"
%%% 919-577-9882 %
%%%% <ya***@ieee.org> % 'Shangri-La', *A New World Record*, ELO
http://home.earthlink.net/~yatescr

[John Harrison]

"Randy Yates" <ya***@ieee.org> wrote in message
news:k6**********@ieee.org...

Thanks David, John, and Mike. I guess I see the point now, but
I don't see why the compiler couldn't have been designed to
forward scan the file and pick up that information without
the user having to go through these hoops - something like
a "two-pass" compiler design.

One of the points of forward references is when the real definition is in a
different file and you don't want the compiler to have to compile the real
definition. This is useful in header files to cut down on dependencies
between header files.

john

[Mike Wahler]

"Randy Yates" <ya***@ieee.org> wrote in message
news:k6**********@ieee.org...

Thanks David, John, and Mike. I guess I see the point now, but
I don't see why the compiler couldn't have been designed to
forward scan the file and pick up that information without
the user having to go through these hoops - something like
a "two-pass" compiler design.

This isn't an issue of compiler design, but language design.

Consider that C++ (like C) is designed to allow separate
compilation of modules of the same ultimate application.
Forward scan *which* file? (the needed one might not, and
often does not, yet exist).

-Mike