Array subscription

[mingke]

Hi....

I have a problem with my declaring my array. I just keep getting message like this:
invalid types `float[9][float]' for array subscript

Here my code:

#define size 2
#define size2 size*size
#define point2 (size+1)*(size+1)
int main ( )
{
float ax[size2][24],ay[size2][24]; 
float R[size2][point2],S[size2][24][point2]; 
float e;
for (i=0; i<size2; i++){
     for (l=0;l<24;l++){
          for (j = 0; i < Lx[i]; j++){
        for (k = 0; k < Ly[i]; k++){
                    e=j+k*Lx[i];
                     R[i][l][e]=ax[i][l];
                      S[i][l][e]=ay[i][l];
                    }
              }
        }
}
}
 

I don't know what's wrong..
if you could help me, I really appreciate it

Thank you

[weaknessforcats]

What is Lx? It's not in your snippet.

[mingke]

Sorry...
I cut my code because it is so long...
Here my complete code (although I cut a few part)

#include <stdlib.h>
#include <math.h>
#include <stdio.h>
#include <conio.h>
 
#define size 2
#define size2 size*size
#define point1 (size+1)
#define point2 (size+1)*(size+1)
#define rotation 45
#define deg rotation*3.14159265/180
 
int main()
{
float xr[point2], yr[point2];
float xb[point2], yb[point2];
float bx[size2][4],by[size2][4];  
float bx_max[size2],bx_min[size2],by_max[size2],by_min[size2];        
float Lx1[size],Lx2[size],Ly1[size],Ly2[size];  
float Lx[size],Ly[size]; 
float R[size2][point2],S[size2][24][point2]; 
int i,j,k,l,flag1,flag2;
int a,d;
 
    for (i = 0; i < point1; i++){
        for (j = 0; j < point1; j++){
           xr[j+i*point1] = -(size/2) + j;
           yr[j+i*point1] = (size/2) - i;
          }
    }
    for (i = 0; i < point2; i++){
         xb[i] = xr[i]*cos(deg) - yr[i]*sin(deg);
           yb[i] = xr[i]*sin(deg) + yr[i]*cos(deg);
    }
for (i = 0; i < size; i++){
  for (j = 0; j < size; j++){            
     a=j+i*size;
     by[a][0]=xr[a+i]*sin(deg) + yr[a+i]*cos(deg);    
     by[a][1]=xr[a+1+i]*sin(deg) + yr[a+1+i]*cos(deg);   
     by[a][2]=xr[a+i+point1]*sin(deg) + yr[a+i+point1]*cos(deg);
     by[a][3]=xr[a+i+1+point1]*sin(deg) + yr[a+i+1+point1]*cos(deg);
     bx[j+i*size][0]=xb[j+i*size+i];
     bx[j+i*size][1]=xb[a+i+1]; 
     bx[j+i*size][2]=xb[a+i+point1];                       
     bx[j+i*size][3]=xb[1+i+a+point1];
   }
}
for (i=0; i<size2; i++){
          for (j = 0; j < 4; j++){
            bx_max[i]=bx[i][0];
                if (bx[i][j]>bx_max[i]){
                    bx_max[i]=bx[i][j];
                }
            by_max[i]=by[i][1];
                if (by[i][j]>by_max[i]){
                    by_max[i]=by[i][j];
                }
       }
  }               
   for (i=0; i<size2; i++){
          for (j = 0; j < 4; j++){
            bx_min[i]=bx[i][0];
            if (bx[i][j]<bx_min[i]){
                bx_min[i]=bx[i][j];
            }
           // by_min[i]=by[i][0];  
            if (by[i][j]<by_min[i]){
                by_min[i]=by[i][j];
            }
       }
   }
for (i=0; i<size2; i++){
       Lx1[i]=ceil (bx_max[i]);
       Lx2[i]=floor (bx_min[i]);
       Ly1[i]=ceil (by_max[i]);
       Ly2[i]=floor(by_min[i]);
       Lx[i]=Lx1[i]-Lx2[i];
       Ly[i]=Ly1[i]-Ly2[i];
  }
for (i=0; i<size2; i++){
      for (l=0;l<24;l++){
        for (j = 0; i < Lx[i]; j++){
          for (k = 0; k < Ly[i]; k++){
               e=j+k*Lx[i];
               R[i][l][e]=ax[i][l];
               S[i][l][e]=ay[i][l];
           }
       }
    }
 }
 
}
 

while ax and ay are values that have been stated before. I'm sorry I didn't put it in my code because it's so long.
And the my problem is it's said that I have invalid types `float[9][float]' for array subscript (R[i][l][e] and S[i][l][e])


Thank you

[newb16]

declaration: float R[size2][point2]
usage: R[i][l][e]
Decalration contains 2 indexes, and you try to index it with 3 indexes.

[mingke]

@newb16

thanks for your respond....

But how about S[i][l][e]?
I declared it with three index and use it with three index too....
Same things happen if I declared R with three index...

Is it possible because e contains Lx[i] and somehow I declared it wrong?

[newb16]

Maybe you can't use float as array index, but only integral type. I tried to compile the code you posted, but got an 'e was not declared in this scope' error at line 84.

[weaknessforcats]

Why are floats being used at all?

I have always recommended that unless you can write down on paper the precise technical reason you need a float you should double instead.

You lose accuracy converting from double to float.

Next, array indexes are integers. When you use a float or a double to calculate an index, all of the integers are converted to double and the calculation is done as a double. Then since array indexes are integers, the resultant double is converted to an int by chopping off the decimals and since a double can hold a bigger value than an int, maybe some of the lower digits are chopped off as well.

You cannot convert a double to an int.

Before I can go further, I would need code that compiles without warnings.

[JosAH]

@weaknessforcats
It's more dramatic than that: if you use floats or doubles for an index expression the code violates a 'shall clause'. This is what the C Standard says:

3.3.2.1 Array subscripting

Constraints

One of the expressions shall have type "pointer to object type ,''
the other expression shall have integral type, and the result has type
"type .''

Violating a shall clause is a crime: error messages must be given and the author of the violation must be defenestrated. This is what the C Standard says about those 'shall clauses':

In this Standard, "shall'' is to be interpreted as a requirement
on an implementation or on a program; conversely, "shall not'' is to
be interpreted as a prohibition.

kind regards,

Jos

[donbock]

It can be more surprising than you expect to invoke "undefined behavior" or "implementation-defined behavior" in your C program.

I don't know if this is just an urban legend, but the word is that the implementation-defined reaction of early versions of gcc was to launch a unix game such as NetHack or Rogue whenever you compiled a program containing #pragma.

The only difference between "undefined behavior" and "implementation-defined behavior" is that the compiler has to document what happens in the latter case. Other than that, anything goes. It would be legal from a C Standard point-of-view to reformat your hard drive.

Defenestration is restrained behavior.

[JosAH]

@donbock
It is true, I saw the sources: when the cppp saw a #pragma it fired up a game. (I forgot which one). It was a single file mess of a program; later versions of the preprocessor have cleaned up a lot.

kind regards,

Jos