"initializer element is not constant"

[mettle]

I'm trying to define a bunch of consant global variables in C, but I'm getting a "initializer element is not constant" error I think because I'm trying to define one in terms of the other.

The code is simply:

int MINFREQ = 200;
int MAXFREQ = 700;
int JND = 5; /* just noticable difference in Hz */
int SEGMENTS = (MAXFREQ-MINFREQ)/JND;

Similarly, I can't seem to define an array using any global variables.
So when I try:

int SEGMENTS = 100;
struct sVowel {
int iF[SEGMENTS];
char sIPA[5];
};

I get:

error: variable-size type declared outside of any functio

Is there a way around this?
I tried defining the first two as constants but that didn't seem to work.

Thanks,
M.E.

[Banfa]

I assume you are using C and not C++ ?

Anyway in both these cases in C marking something as const does really achieve anything. These are not situations where you can use a variable, global variable initialisation or array size definition.

About your only choice is to use #define instead

#define MINFREQ     200
#define MAXFREQ     700
#define JND         5 /* just noticable difference in Hz */
int SEGMENTS = (MAXFREQ-MINFREQ)/JND;
 

#define SEGMENTS 100
 
struct sVowel {
    int iF[SEGMENTS];
    char sIPA[5];
};
 

[mettle]

Thank you for your help.

Just out of curiosity, how does C know whether MINFREQ is an int or a double?
If I had defined it was 200.0 would it be a double instead of an int?

I assume you are using C and not C++ ?

Anyway in both these cases in C marking something as const does really achieve anything. These are not situations where you can use a variable, global variable initialisation or array size definition.

About your only choice is to use #define instead

#define MINFREQ     200
#define MAXFREQ     700
#define JND         5 /* just noticable difference in Hz */
int SEGMENTS = (MAXFREQ-MINFREQ)/JND;
 

#define SEGMENTS 100
 
struct sVowel {
    int iF[SEGMENTS];
    char sIPA[5];
};
 

[Banfa]

Just out of curiosity, how does C know whether MINFREQ is an int or a double?
If I had defined it was 200.0 would it be a double instead of an int?

Yes (to the second question).

basically if you put in a decimal point if is a double, otherwise it is an int. You can modify the type of a constant using the letters U (unsigned), L (long) and F (FLOAT) so

200         // int constant
200U        // unsigned int constant
200L        // long constant
200UL       // unsigned long constant
200.0       // double constant
200.0F      // float constant