Thread: Getting segmentation fault with strtok()

I assume that before the first strtok call, that you have actually called it with a different string, right? Because you're not actually doing that anywhere.


Quzah.

What about if the return value of strtok() is NULL (which it uses to indicate that no matching tokens were found)?

[edit] Oh, and as quzah said, the first time you call strtok() you're supposed to pass the string you want it to tokenize . . . . [/edit]

What's the definition of hold.

You don't check curr for NULL, so if you ever end up back at the strtok with no comma, the line that checks for " will fail.

And I personally prefer a solution that does something like this:

Code:

look at current character
if quote, then copy until next quote
else if comma, save current string, set string to empty
else add char to current string.

All that strtok/strcat/strcpy seems awfully complicated for such a simple task.

--
Mats

Originally Posted by Albinoswordfish

Code:

  printf("testing\n");

It may (or may not) make a difference in your debugging here, but for future reference *always* follow a debugging printf() with fflush(stdout):

Code:

printf("testing\n"); fflush(stdout);

This is because the stdout buffer is probably not flushed just by a '\n'; it may be held longer than than that. So, if you are moving that line around, or using several such printf() lines to isolate a seg fault, you *can* be misled as to where it occurs if you do not use fflush(stdout).

I was under the impression that stdout was automatically flushed by a '\n' when stdout refers to a terminal. I'm not sure how portable this is, though, and a fflush() never hurts (unless you do it on stdin...).