Thread: Suggestions for improving.

I think matsp's statement was referring to how arrays and pointers have some semantic differences. For example, you can't increment arrays.

Writing to read-only memory = Seg fault + crash

Not necessarily. In programs compiled with Turbo C, when you dereference NULL, you get a message like "NULL pointer dereference" or something similar, and the program continues executing. You'll usually get a segmentation fault, but not all of the time.

[edit] Ah yes, it was "NULL pointer assignment".

But they are essentially the same because they simply point to allocated memory.

Pointers are a little more flexible on this point, because you can have pointers pointing to the middle of arrays or dynamically allocated memory chunks. [/edit]

At the hardware level and to the computer, arrays and pointers are more or less the same thing: contiguous blocks of memory that we reference usually by the address of the first element.

Now in C, there are times that arrays and pointers are considered explicitly different, at least to the compiler. You can say they are the same, but the compiler will not always agree.

TBH, if you understand why the compiler is throwing a fit and you understand the true difference between arrays and pointers, to interchange them when you know what you're doing isn't really a problem. Again, it's all provided you know what's going on.

Originally Posted by IceDane

What arrays are, basically, are simplified versions of pointers.

No, that's not correct. The C-compiler can make a pointer from an array, and you can use array syntax to access elements from a pointer, but an array and a pointer are semantically different, even if there is little difference in syntax. As I said earlier, an array declaration forms a piece of memory that the program can use to store/retrieve data from. A pointer contains some memory address - which of course could be an array, like in your example below - the ptr variable is set to the address of array (same as &array[0] - the compiler essentially allows this shortened form, because it's a very common way to do things).

Code:

#include <stdio.h>

int main(void)
{
    char array[] = {"String"}, *ptr = array;
    
    printf("*(ptr+3) = %c\n"
           "array[3] = %c", *(ptr+3), array[3]);
           
    return 0;
}

We both know what that code prints. Like the above poster said, on hardware level, both are just contiguous blocks of memory in the program's address space - It's only C which differentiates.

Of course, everything is memory in a computer, and it's just a question of how you use the memory. But if we look at it from a assembler standpoint, you have

Code:

array: db "String",0
pptr:   dw array    // renamed to make it easier to follow, since ptr is a keyword in MASM
fmt:    db "*(ptr+3) = %c\narray[3] = %c", 0
...
     mov eax, offset array
     push byte ptr [eax+3]
     mov eax, dword ptr pptr
     push byte ptr [eax+3]
     push offset fmt
     call   printf

Now, that's not exactly the same thing, is it? It just gives the same result.

--
Mats

http://c-faq.com/aryptr/aryptr2.html

Originally Posted by IceDane

If you take a pointer and an array, and print their hexadecimal values, you'll get addresses for both. Why? Because the first element of the array is simply the address of the allocated memory, and subsequent elements are simply the memory address incremented respectively. Arrays are simply an attempt to simplify string handling in C - They are still exactly the same as pointers when it comes to memory.

No, arrays are a way to statically (or semi-statically, in the case of local variables inside functions) to allocate memory for something. Pointers are a way to refer to things that are allocated in memory. There is a difference.

By the way, arrays are used for a lot of things that aren't strings, although I would agree that strings are the most common case. But you can have arrays of structs, classes (assuming C++), and of course multidimensional arrays.

I think you have it back to front - pointers are made to look like arrays, because it makes life easier dealing with pointers that way.

--
Mats

> Pointers can substitute arrays in any situation,
How wrong can you be?

Code:

int main() {
    int arr[] = { 1, 2, 3, 4, 5 };
    int *p = { 1, 2, 3, 4, 5 };
    return 0;
}

The second one will get you a face-full of errors.
Read the C.L.C FAQ link Dave pointed out to you.

> The C-compiler can make a pointer from an array
Except in some special cases - the first being the one I've just posted.

The second is when you use sizeof().
sizeof(array) and sizeof(ptr) give very different answers.

There is a 3rd exception, which I'll leave you to discover by reading the FAQs

Looks like the topic was very essential to be opened.I am happy for that. I can make a statement too I think.There is a difference. Because arrays dont really act like a pointer.

Code:

#include <stdio.h>

int main(void)
{
 char a[6]="naber";
	a=&a[3];
	printf("&#37;s",a);
    return 0;
}

This code gives error. Unlike :

Code:

#include <stdio.h>

int main(void)
{
 char *a="naber";
	a=&a[3];
	printf("%s",a);
    return 0;
}

This code however, does not give error. In my opinion , the arrays just act in resemblence with pointers. Because in order to change the adress of string (assume it is an array) we have to use it as &string[0]. I may be looking to the case simply. But I have been thinking this way. If I am wrong please clarify me. By the way , here is the code of the program which counts each different word in a sentence:

Code:

#include <stdio.h>
#include <string.h>
void sayac(char *,char *,int,int *);
int main()
{
struct str {
		char d[10];
	} st[8];
char s[80];
char kopya[80];
	gets(s);
	sprintf(kopya,"%s",s);
	char *b;
	b=strtok(s," ");
	int x=0;
	int i=0;
	for(i=0;i<8;i++) {
	
          x+=1;
		sprintf(st[i].d,"%s",b);
				b=strtok(NULL," ");
		
		if(b==NULL) break; }
  int y=0;
	int k=0;
	int control=0;
	int p=0;
    for(i=0;i<x;i++) {
    for(k=0;k<i;k++)
			if((strcmp(st[i].d,st[k].d))==0) control=1;
		if(control==1) continue;
 y=strlen(st[i].d);
		sprintf(s,"%s",kopya);
		
		sayac(s,&(st[i].d[0]),y,&p);
              
		printf("%s den %d tane var\n",st[i].d,p);
 p=0;     
		
		}
return 0;
}           

		

	void sayac(char *ana,char *str,int t,int *ptr) {
		if(strstr(ana,str)==NULL) return;
		else {
			ana=strstr(ana,str);
			*ptr=*ptr+1;
			sayac(&ana[t],str,t,ptr); }
	}

Sorry for the mess of code lines. This is the way I write.

Thank you for every response. =) By the way that post someone posted which leads us to another site , actually explains what I knew. That is true isnt that?

A few improvements, mostly style, but also efficiency.

Code:

    for(i=0;i<x;i++) {
    for(k=0;k<i;k++)
			if((strcmp(st[i].d,st[k].d))==0) control=1;
		if(control==1) continue;
if (!control) {
 y=strlen(st[i].d);
		sprintf(s,"%s",kopya);
		
		sayac(s,&(st[i].d[0]),y,&p);
              
		printf("%s den %d tane var\n",st[i].d,p);
 p=0;    
} 
		
		}

Once control is one, it will stay one for forever. I suspect you want to reset it to zero someplace.

Code:

         int y = 0;   // Remove the whole line. 
....
         sayac(s,&(st[i].d[0]),&p);   // Remove y
....
	void sayac(char *ana,char *str,int *ptr) {
		ana=strstr(ana,str);
		if(ana) {
			(*ptr)++;
			sayac(ana,str,ptr); 
                }
	}

This avoids calling strstr twice for no reason at all, and simplifies the if-statement (removes the else-part). the variable t is not used, (y is never changed in the calling code, and y is set to zero, so no reason to pass it.



--
Mats

Originally Posted by ozumsafa

I dont know if I misunderstood but if you do that , you will always find the same str in ana string.

Ok, if you say so. You still shouldn't have to do strstr twice, which ever way you do it.

Yes I wanted to make control zero , but I copy pasted the codes. Sorry for that.

And PLEASEEE , I know getss fgetss FAQQ!!!

So, if you know how to do it correctly, why isn't your code showing that.

And it would be easier for us to understand what's going on if:
1. The language of printf() and such is English - I know, it's not your native language, neither is it mine (although I do admittedly live in England, and have done so for several years).
2. If you make sure the indentation is correct when posted - it looks like your code has a mixture of spaces and tabs. Tabs aren't necessarily the same width in all applications, so it's best if you can replace them with spaces. [Most editors have a way to replace tabs with spaces all automaticaly].

I presume you do want to learn from this.

--
Mats