Thread: string reversal

Originally Posted by WatchTower

You need return 0 in your main function.

It's optional according to latest C standards.

Originally Posted by Adak

Perhaps it doesn't run on your compiler, but it runs just fine on mine (Turbo C).

Might be because it's an old compiler.

Originally Posted by Adak

Perhaps it doesn't run on your compiler, but it runs just fine on mine (Turbo C).

I am curious to know just what you mean by "run on your compiler". Your code compiles on the MinGW port of gcc 3.4.5 with just one warning (you forgot to #include <string.h>), but it is logically incorrect since it attempts to modify a string literal, as that part of the OP's code was not fixed. My guess is that you (incorrectly) interpreted "even the code you gave is not running" as "even the code you gave does not compile". If not, and you actually ran the program and it worked, then it is just an effect of undefined behaviour - that it compiles is no guarantee of correctness.

Originally Posted by laserlight

I am curious to know just what you mean by "run on your compiler". Your code compiles on the MinGW port of gcc 3.4.5 with just one warning (you forgot to #include <string.h>), but it is logically incorrect since it attempts to modify a string literal, as that part of the OP's code was not fixed. My guess is that you (incorrectly) interpreted "even the code you gave is not running" as "even the code you gave does not compile". If not, and you actually ran the program and it worked, then it is just an effect of undefined behaviour - that it compiles is no guarantee of correctness.

Now after searching the net I get to know that string literals can't be modified as they ar stored in read-only section of the memory by the OS. But my question is why string literals are given such type of a special "treatment"?
And by the way is anything between double quotes a string literal?

Originally Posted by Adak

I mean the compiler is able to run through it's work, and create an executable file, without warnings or errors. Turbo C has a smart compiler which is able to include strlen() functionality, into a program, despite not explicitly including string.h.

The resulting program runs correctly. Some sample output appropriate for July 4th:

Maybe the last line of this FAQ is helpful here.
C-FAQ

Originally Posted by roaan

Hi,

I have a doubt about string reversal. The code below reverses the string character by character but when i give it to print out the result there occurs an error. I am unable to figure out where am i going wrong.

insert

Code:

#include <stdio.h>
#include <stdlib.h>


void xstrrev(char *);

int main(){

	int i;
	char *s[] = {
		"To err is human...",
		"But to really mess things up...",
		"One needs to know C!!"
	};

	for(i=0;i<3;i++){
		xstrrev(s[i]);
	}
}


void xstrrev(char *a){


	int len,j;
	char *p;

	printf("\n\r%s", a);
	len = strlen(a);
	printf("\n\r%d", len);
	p = malloc(len+1);

	for(j = 0;j<=len;j++){	
		p[j] = a[len-j];
		printf("\n\r%c%c", a[len-j], p[j]);
		if(j == len)
			printf("\n\r%s", p);
	} }

After going through the details of arrays of pointers and character strings(and also hijacking your thread) I made 3 solutions of this question. But in all of them the main thing is that I'm swapping the characters of each string until the mid-character. Out of the three approaches I took one of them was with malloc which is similar to your code posted. But I dont understand what you're trying to do in the colored lines. What I've done is after assigning the storage to p, copied a into p and then the same swapping thing. This way the swapped string gets stored in p.

Originally Posted by Adak

I mean the compiler is able to run through it's work, and create an executable file, without warnings or errors.

That is, the compiler was able to compile the code without warnings or errors. That is different from saying that the program does not run (but admittedly I see the potential for confusion since a program that does not compile obviously does not run: perhaps BEN10 should have mentioned a run time error or something).

Originally Posted by Adak

Turbo C has a smart compiler which is able to include strlen() functionality, into a program, despite not explicitly including string.h.

Haha. That is not the sign of a "smart compiler", since a compiler (or linker/loader invoked by the compiler) might link to the C standard library by default, and it is possible for a function to be used without being declared at the point of usage. One possible reason that you did not get a warning is that <string.h> could have been included by a header that you did include - relying on this would be relying on something implementation specific. Another possible reason is that the warning level was not high enough.

Originally Posted by Adak

The resulting program runs correctly. Some sample output appropriate for July 4th:

As such, you are observing the effect of undefined behaviour. It appears to work in your case, but is nonetheless incorrect.

Originally Posted by BEN10

But my question is why string literals are given such type of a special "treatment"?

I believe there are practical reasons related to how implementations might implement string literals, but consider this: what does it mean to modify a literal? For example, does this make sense to you?

Code:

1 = 2;

Thanks everybody. Now I think some of the loops in my mind are opening regarding pointers.