strcat usage with char* and char []

fjkum · 12-12-2008, 11:59 PM

Hi,

I wrote a simple program in C under linux but now it segfault on me when I run it, below is a snippet of the code:

char *fullstr;
char *pStr = "second";
char arStr[10] = "third";

if( pStr != NULL )
{
strcat( fullstr, pStr );
strcat( fullstr, arStr ); <--- this causes segfault
}

Any idea?

elprawn · 12-13-2008, 08:37 AM

Try this...

Code:

char *fullstr;
fullstr = new char[strlen(pStr) + strlen(arStr)];
strcpy(fullstr, pStr);
strcat(fullstr, arStr);

mallux · 12-13-2008, 02:32 PM

The reason that your program is segfaulting is because you haven't allocated any memory for "fullstr". On my system your program (quite rightly) falls over at the first strcat rather than the second.

elprawn's example will work, but you can only use "new" in C++ rather than C. In C you would have to use malloc (or one of its related cousins such as calloc etc), thus:

Code:

fullstr = malloc(strlen(pStr)+strlen(arStr));

Don't forget to free the memory when you've finished with it:

Code:

free(fullstr);

elprawn · 12-13-2008, 05:25 PM

Ah yeah in C. Sorry!

ta0kira · 12-13-2008, 07:46 PM

!= NULL isn't ever necessary except for clarity (unless you have a class in C++ without an operator bool or similar.)

Also:

Code:

fullstr = calloc(strlen(pStr)+strlen(arStr)+1);

Or:

Code:

fullstr = malloc(strlen(pStr)+strlen(arStr)+1);
fullstr[0] = 0x00;
//or...
memcpy(fullstr, pStr, strlen(pStr) + 1);

strcat looks for the end of the first string; it doesn't automatically start from the beginning for the first call to it. You have a 1-(255/256)^n chance of not writing out of bounds with the first call to it, where n is the size of the second string + 1, if you don't make sure a null-character is in place.
ta0kira

ErV · 12-13-2008, 10:35 PM

Quote:

Originally Posted by elprawn

Code:

fullstr = new char[strlen(pStr) + strlen(arStr)];

Incorrect.
Where is the extra byte for trailing zero?

ta0kira · 12-13-2008, 11:11 PM

Quote:

Originally Posted by ErV

Incorrect.
Where is the extra byte for trailing zero?

Already got that one.
ta0kira

elprawn · 12-13-2008, 11:12 PM

Quote:

Originally Posted by ErV

Incorrect.
Where is the extra byte for trailing zero?

Are you talking about null termination? I realised this afterwards but someone else already pointed it out.