Recusive strcpy

[justdoit22]

Need help in Recusion

char* StrCpy(char* output, char* input)
{
	
	if (*input != NULL)
	{
		*output = *input;
		*output++;
		*input++;

		StrCpy(output,input);
	}
	else *output= NULL;
	return output;
}

If i Give input as "ASD"
output is "SD"
where is the problem..
need quick help

[Salem]

Perhaps because you return the value of output AFTER you have incremented it.

[Dave Evans]

It works for me. How are you defining your actual parameters input and output in the program that calls it?

Here's mine:

#include <iostream>

using namespace std;

int main()
{
  char* StrCpy(char* output, char* input);

  char outstr[100];
  //char instr[100] = "This is a test"; // try this : still works
  char instr[100] = "ASD";

  cout << "instr  = <" << instr << ">" << endl;
  StrCpy(outstr, instr);
  cout << "outstr = <" << outstr << ">" << endl;

  cout << endl << "Now, you enter a word to copy: ";
  cin  >> instr;

  cout << "instr  = <" << instr << ">" << endl;
  StrCpy(outstr, instr);
  cout << "outstr = <" << outstr << ">" << endl;

  return 0;
}


char* StrCpy(char* output, char* input)
{
    
    if (*input != NULL)
    {
        *output = *input;
        *output++;
        *input++;

        StrCpy(output,input);
    }
    else *output= NULL;
    return output;
}

Tested on Windows XP with Borland bcc, Microsoft Visual C++, and GNU g++ (Cygwin).

Tested on Linux with GNU g++.



Dave

[Prelude]

>It works for me.
You must not have been paying attention to the output. I got the same problem as the OP.

>where is the problem..
You have several. The first is that NULL and '\0' may not evaluate to the same thing. In C++ they probably will, but since NULL is used in pointer context and not character context, you should avoid using it as you were. Next, to be strictly correct, input should be made const because you never change the contents of the string. And lastly, your actual problem. It is as Salem said. If you refrain from using ++ in your recursive functions in favor of adding 1 to the recursive call, you can avoid problems like these. Here are my changes:

char* StrCpy(char* output, const char* input)
{
  if (*input != '\0')
  {
    *output = *input;
    StrCpy(output + 1,input + 1);
  }
  else
    *output = '\0';

  return output;
}

[Dave Evans]

With all due respect to your expertise, I repeat that "it works for me". I submitted a complete program using his test string, and ran the thing on the compilers that I indicated. My point was, I was not trying to get him to improve his style just yet; I was trying to understand what his problem was so that he could fix it.

(If it ain't broke, you can't fix it.)

Will you post a complete program that breaks with ++ but works the other way? I would like to run it through my compilers. I am very interested in finding things that work in one programming environment that break in others, and understanding the differences.

Note a confusion factor is in his routine where he has

*input++;
*output++;

It would undoubtedly make more sense to have

input++;
output++

These have the same effect (incrementing local copies of the variables input and output).

I didn't change this in my response to him, since it had nothing to do with causing a problem.

As for your points:

Yes, I know that NULL is a pointer type, and, in fact g++ gave me the warning with his original code. I changed it to '\0' to get rid of the warning. This had nothing to do with his problem, so I changed it back to duplicate his code exactly before running the test with different input strings.

As far as making the input string const: that's OK, and probably good programming practice, but had nothing to do with his problem.

At a given level, calling a function with a value of (input+1) is no different than incrementing the variable (input), then calling with this incremented value.

Dave

[Salem]

I rather suspect the difference the OP sees is down to

StrCpy(outstr, instr);
cout << "outstr = <" << outstr << ">" << endl;

Vs.

cout << "outstr = <" << StrCpy(outstr, instr) << ">" << endl;

The former will work whether you mess up the increment or not, but the latter will show the off-by-one error.

[Dave Evans]

Ahhh, now I get it. This is kind of important. That's why I asked for an example.

When somebody tells me something doesn't work, I need to know how he tested it.

I completely missed the point that the return value of the function was wrong. I was paying attention to he actual copying (which worked OK), but there was a real bug in the function, which has now been fixed.

On the other hand, when I submit a test program and someone tell me, "You must not have been paying attention to the output," it's not very helpful. My program gave the expected output, but wasn't an adequate test of the function in question.

(You can't prove a program is correct by testing.)

Thanks a lot.

Dave