RE: How to pass 2D variable-sized arrays in C++?

RE: How to pass 2D variable-sized arrays in C++?

[Quang Nguyen (Ngo)]

Hi Jason,

Try this:

void func(int c, int r, float f[][])
        {
        printf("hello\n");
        }

main()
        {
        int c = 5, r = 7;
        float f[r][c];

        func(c, r, f);
        }

--
Quang



-----Original Message-----
From: Jason Mancini [mailto:jayrusman@hotmail.com] 
Sent: Wednesday, September 25, 2002 7:37 PM
To: gcc-help@gcc.gnu.org
Subject: How to pass 2D variable-sized arrays in C++?



Hello,
How does one type the 2D variable size array "f" below?
I've tried nasty casts that all get rejected, cheating
with void, etc, etc.  Very stumped!  Sure, I could back
up to a real float** and manually new[] and pass float**,
but that takes all the fun out!

Thanks,
Jason


void function(int c, int r, ??? f )
{
}

main() {
  int c(5), r(7);
  float f[r][c];
  function(c, r, f);
};


--


_________________________________________________________________
Join the world's largest e-mail service with MSN Hotmail. 
http://www.hotmail.com


________________________________________________________________________
This email has been scanned for all viruses by the MessageLabs SkyScan
service. For more information on a proactive anti-virus service working
around the clock, around the globe, visit http://www.messagelabs.com
________________________________________________________________________

________________________________________________________________________
This email has been scanned for all viruses by the MessageLabs SkyScan
service. For more information on a proactive anti-virus service working
around the clock, around the globe, visit http://www.messagelabs.com
________________________________________________________________________

OT: RE: How to pass 2D variable-sized arrays in C++?

[Claudio Bley]

>>>>> "Quang" == Quang Nguyen (Ngo) <quang.nguyen@tapeware.com> writes:

    Quang> Hi Jason, Try this:

    Quang> void func(int c, int r, float f[][]) { printf("hello\n"); }

    Quang> main() { int c = 5, r = 7; float f[r][c];

    Quang>         func(c, r, f); }

Note, that this is no valid C++ code. You may just leave out the first
dimension of an multidimensional array ( e.g. f[5][] or f[5][4][8][] )
because the compiler needs this information in order to find the
position of an element.

The second problem is that it might not do what you expect when you
want to use it in the usual notation f[i][j] which is actually 
equivalent to (*(*(f+i)+j)).

The correct solution is to use it like so:

void print_mn (int m, int n, float *f) 
{
	for (int i = 0; i < m; ++i) {
		for (int j = 0; j < n; ++j) {
			cout << f[i*n+j] << '\t';
		}
		cout << endl;
	}
}

float[4][5] f;
print_mn (4, 5, &f[0][0]);


To the OP: Please post this sort of language specific (non GCC
specific) questions in the appropriate forum next time. And may I
suggest that you get a good book on C++ programming?

-- 
Claudio Bley                                 ASCII ribbon campaign (")
Debian GNU/Linux advocate                     - against HTML email  X 
http://www.cs.uni-magdeburg.de/~bley/                     & vCards / \

Re: How to pass 2D variable-sized arrays in C++?

[Claudio Bley]

>>>>> "John" == John Love-Jensen <eljay@adobe.com> writes:

    >> OK, but what's your point here? We were talking about
    >> dynamically sized 2D arrays where the dimension of the array
    >> passed to `foo' is not known at compile time.

    John> In that case, I refer back to my first answer: typedef
    John> vector<vector<float> > myarray;

    John> void foo(myarray& a) ...

Of course, I'd prefer that too - but I guess you don't always have the
choice.

-- 
Claudio Bley                                 ASCII ribbon campaign (")
Debian GNU/Linux advocate                     - against HTML email  X 
http://www.cs.uni-magdeburg.de/~bley/                     & vCards / \

Re: How to pass 2D variable-sized arrays in C++?

[John Love-Jensen]

> OK, but what's your point here? We were talking about dynamically
> sized 2D arrays where the dimension of the array passed to `foo' is
> not known at compile time.

In that case, I refer back to my first answer:
typedef vector<vector<float> > myarray;

void foo(myarray& a) ...

--Eljay

Re: How to pass 2D variable-sized arrays in C++?

[Claudio Bley]

Hi,

>>>>> "John" == John Love-Jensen <eljay@adobe.com> writes:

    John> Hi Jason, Try this out for size:

OK, but what's your point here? We were talking about dynamically
sized 2D arrays where the dimension of the array passed to `foo' is
not known at compile time.

-- 
Claudio Bley                                 ASCII ribbon campaign (")
Debian GNU/Linux advocate                     - against HTML email  X 
http://www.cs.uni-magdeburg.de/~bley/                     & vCards / \

Re: How to pass 2D variable-sized arrays in C++?

[John Love-Jensen]

Hi Jason,

Try this out for size:

- - - - - - - - - -
#include <iostream>
using namespace std;

typedef float myarray[3][7];

void foo(myarray& a)
{
    for(int i = 0; i < 3; ++i)
        for(int j = 0; j < 7; ++j)
            cout << a[i][j] << (j == 6 ? "\n" : " ");
}

int main()
{
    myarray a;
    a[0][0] = 0.0f;
    a[0][1] = 0.1f;
    a[0][2] = 0.2f;
    a[0][3] = 0.3f;
    a[0][4] = 0.4f;
    a[0][5] = 0.5f;
    a[0][6] = 0.6f;
    a[1][0] = 1.0f;
    a[1][1] = 1.1f;
    a[1][2] = 1.2f;
    a[1][3] = 1.3f;
    a[1][4] = 1.4f;
    a[1][5] = 1.5f;
    a[1][6] = 1.6f;
    a[2][0] = 2.0f;
    a[2][1] = 2.1f;
    a[2][2] = 2.2f;
    a[2][3] = 2.3f;
    a[2][4] = 2.4f;
    a[2][5] = 2.5f;
    a[2][6] = 2.6f;
    foo(a);
}
- - - - - - - - - -

--Eljay

RE: How to pass 2D variable-sized arrays in C++?

[Claudio Bley]

>>>>> "Jason" == Jason Mancini <jayrusman@hotmail.com> writes:

    Jason> Hi again, After many tries, I propose that this is the
    Jason> "correct solution".  The compiler should hide the address
    Jason> calculation for me -- this is why we have compilers after
    Jason> all!  Thanks, Jason Mancini


    Jason> void func(int c, int r, float *fa) { 
    Jason>    float (*f)[c] = (float (*)[c])fa;
    Jason>    ... f[xr][xc] ... }

Yeah, this is cool. I didn't know this is possible. The drawback is
that it might not work with any other compiler out there except
gcc. (I can confirm it doesn't work with Intel's C++ compiler for
example)

I myself wouldn't use such esoteric features (if it really is one -
maybe Intel's compiler is just faulty?). I would like to see some
reference to the standard spec where this is defined.

Cheers
-- 
Claudio Bley                                 ASCII ribbon campaign (")
Debian GNU/Linux advocate                     - against HTML email  X 
http://www.cs.uni-magdeburg.de/~bley/                     & vCards / \

RE: How to pass 2D variable-sized arrays in C++?

[Jason Mancini]

Hi again,
After many tries, I propose that this is the "correct solution".
The compiler should hide the address calculation for me -- this
is why we have compilers after all!
Thanks,
Jason Mancini


void func(int c, int r, float *fa)
{
  float (*f)[c] = (float (*)[c])fa;
  ... f[xr][xc] ...
}

main()
{
  int c = 5, r = 4;
  float f[r][c];
  func(c, r, &f[0][0]);
}


-------------------------------

Claudio Bley wrote:

The second problem is that it might not do what you expect when you
want to use it in the usual notation f[i][j] which is actually
equivalent to (*(*(f+i)+j)).

The correct solution is to use it like so:

void print_mn (int m, int n, float *f)
{
for (int i = 0; i < m; ++i) {
for (int j = 0; j < n; ++j) {
cout << f[i*n+j] << '\t';
}
cout << endl;
}
}

float[4][5] f;
print_mn (4, 5, &f[0][0]);



_________________________________________________________________
Send and receive Hotmail on your mobile device: http://mobile.msn.com

RE: How to pass 2D variable-sized arrays in C++?

[Jason Mancini]

Hello,
Thank you for the thorough explanation!  That is a sol'n I
thought of, but I wasn't sure it could be applied to runtime
dynamic-sized arrays.
Thanks,
Jason Mancini

-------------------------------

Claudio Bley wrote:

The second problem is that it might not do what you expect when you
want to use it in the usual notation f[i][j] which is actually
equivalent to (*(*(f+i)+j)).

The correct solution is to use it like so:

void print_mn (int m, int n, float *f)
{
	for (int i = 0; i < m; ++i) {
		for (int j = 0; j < n; ++j) {
			cout << f[i*n+j] << '\t';
		}
		cout << endl;
	}
}

float[4][5] f;
print_mn (4, 5, &f[0][0]);


_________________________________________________________________
Join the worldÂ’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com

How to pass 2D variable-sized arrays in C++?

[Jason Mancini]

Hello,
How does one type the 2D variable size array "f" below?
I've tried nasty casts that all get rejected, cheating
with void, etc, etc.  Very stumped!  Sure, I could back
up to a real float** and manually new[] and pass float**,
but that takes all the fun out!

Thanks,
Jason


void function(int c, int r, ??? f )
{
}

main() {
  int c(5), r(7);
  float f[r][c];
  function(c, r, f);
};


--


_________________________________________________________________
Join the worldÂ’s largest e-mail service with MSN Hotmail. 
http://www.hotmail.com