Thread: replace character while writing into a file

Code:

#include <stdio.h>
#include <string.h>

int main(int argc, char **argv)
{
   FILE *fin, *fout;
   char *s;
   fin = (argc > 1) ?fopen(*++argv, "r"):stdin;
   fout = (argc > 2)?fopen(*++argv, "w"):stdout;
   if (fout && fin) 
      do { 
          char buffer[1000]; 
          char *p;
          while(s = fgets(buffer, sizeof(buffer), fin))
          {
              p = s;
              while(p = strchr(s, 'w'))
              {
                  s = p;
                  if (p - buffer > 3 && p[-1] == 'o' && p[-2] == 'l' && p[-3] == 'f' && 
                      p[1] == 'o' && p[2] == 'r' && (p[1] = 'e'));
               }
               fputs(buffer, fout);
          }
        } while(s);
    else
       return 1;
   return 0;
}

--
Mats

Isn't a file descriptor of stdin going to be 0, and therefore, your AND will fail if the program is run with stdin redirection?

Originally Posted by Dino

Isn't a file descriptor of stdin going to be 0, and therefore, your AND will fail if the program is run with stdin redirection?

But it's not a file-descriptor, it's a file-pointer - which is guaranteed to not be 0.

How did you like it otherwise?

--
Mats

I kind of agree with shoutatchickens: if you know that you are going to write "flower" to the second file, why bother reading from the first file?

btw, matsp, this is the C++ programming forum, and it turns out that due to typos your code is neither valid C nor C++, heheh.

Quite clever. I really liked the optional parms and the negative indexing.

Originally Posted by laserlight

I kind of agree with shoutatchickens: if you know that you are going to write "flower" to the second file, why bother reading from the first file?

btw, matsp, this is the C++ programming forum, and it turns out that due to typos your code is neither valid C nor C++, heheh.

Ok, I have updated it to valid C - it won't pass as C++, and I can't be bothered to rewrite it as C++, to be honest.

--
Mats

Actually, there is a logic bug. If an instance of flower is split across buffers, due to really long lines, it won't be found.

Originally Posted by Dino

Actually, there is a logic bug. If an instance of flower is split across buffers, due to really long lines, it won't be found.

Ah, well, can't get everything for free - I suppose we could add:

Code:

{
   int len = strlen(s);
   if (s[len-1] != '\n') fprintf(stderr, "Lines longer than %s characters... Output may be incorrect\n", sizeof(buffer)-1);
}

--
Mats

Yep., or this too:

Code:

if (s[len-1] != '\n') fprintf(stderr, "Lines longer than %s characters... Output may be incorrect (or this program might seg fault ;) )\n", sizeof(buffer)-1);

Good job otherwise. I just notice the assignment of 'e' as part the &&'ed logic. Slick.

I take it you intended to say [strlen(s)-1] instead of [len-1]?

--
Mats

I just copied your code, big boy!!

Originally Posted by Dino

I just copied your code, big boy!!

Oh, I see what you did. It probably won't seg-fault tho' - unlikely that the offset on the stack segfaults.

--
Mats

Yes, you are correct - modern stacks are big.

When I program for my day job, on mainframes, memory allocations are designed (obviously) around page boundaries (4K). If you request 8 bytes, and your request is the first request, or the first request that requires a new page frame, then you get the last 8 bytes in a page. If you touch one byte later, and you don't own that page, you get a fault.

And, I work in assembler, so the concept of a system provided stack is also moot - you get nothing, so it's all up to the programmer (as I assume it is for asm programmers on any platform)