Thread: concatenating a string

I'm not sure why you want to do this this way, but here are the problems I found:

Code:

*(C+j) = (char*) malloc(sizeof(char*));

You are not adding a pointer, you are adding a single char, so get rid of the asterisks.

Code:

*C = updateChar(C,'o');

updateChar operates on C so you do not need (and in this case should not) use the return value. (Just call the function, "updateChar(C,'o');")

Code:

} int j = i++;

you've gone too far. Remove the ++, make those other changes, and the program will print out "Ciaoo"

Originally Posted by MK27

updateChar operates on C so you do not need (and in this case should not) use the return value. (Just call the function, "updateChar(C,'o');")

Not really, since updateChar used without a return value couldn't actually change the size of the array unless you change the values you pass it. Therefore, using the parameters it currently has, there is no way to make that string actually longer without returning a new string. Furthermore, with its current prototype, unless you free the old string outside of the function, you're going to make a memory leak.


Quzah.

Originally Posted by quzah

Not really

In fact really, since I compiled the code with the changes I mentioned (no return value), used printf back in main, and C is now "Ciaoo".

However, I don't think a sane person should write code this way anyway.

Originally Posted by MK27

In fact really, since I compiled the code with the changes I mentioned (no return value), used printf back in main, and C is now "Ciaoo"

Honestly, that doesn't mean a damn thing.

Code:

#include<stdio.h>
int main( void )
{
    char buf[ 2 ];
    buf[ 0 ] = 'H';
    buf[ 1 ] = 'i';
    buf[ 2 ] = '!';
    buf[ 3 ] = '\0';

    printf( "%s\n", buf );
    return 0;
}

This will almost NEVER crash. It doesn't make the code good, correct, or right in any way shape or form.


Quzah.

Okay, I totally agree. Like I said, no one should be writing code like this anyhow, I presume it is some kind of experiment.

I had thought you meant that without the return value, "updateChar" COULD NOT affect "C" globally, which is another issue but not part of this thread.

No I meant this:

Code:

void myfailuremalloc( char * s, size_t t )
{
    s = malloc( t );
}

This won't actually make s point to the string malloc is making once this function ends. Which is what they're trying to do...


Quzah.

Originally Posted by quzah

No I meant this

Ah! Well so That Issue (qv.) is part of this thread then. You're right. However, I noticed testing the OP that using the return value with the original "C" dereferenced caused more some unwelldefined problems. But I shouldn't have recommended what I did considering the malloc() in updateChar. Sorry sorry sorry. A few days of .pl and .js and I just lose my head.

IMO, even if you didn't want to use some standard C functions to do this stuff, this whole approach is even worse than the "worst" scenario I might have dreamt up as humour, I have no idea how this person got on this track or why they think it is even somewhere to stay temporarily. We are for sure ignoring some compiler warnings, methinks.