Question

I am getting a junk character to be output at the very end of some text that I read in:

hum 1345342342 ~Users/Documents ecabd459 //line that was read in from stdin
event action: hum_?
event timestamp: 1345342342
event path: ~Users/Documents
event hash: ecabd459

At the end of the event action value there is a ‘_?’ garbage character that is output as well. That can be rectified by setting the variable’s last position to the null terminator (event.action[3] = '\0') which is all well and good, but I am perplexed by the fact that the other char array event.hash does not exhibit this type of behavior. I am creating/printing them in an identical manner, yet hash does not behave the same.

Note: I was considering maybe this was due to the hash value being followed strictly by a newline character(which I get rid of by the way), so I tested my program with re-ordered input to no avail (that is, added an additional space and word after the hash value’s position on the line).

The relevant code is below:

struct Event{
    char action[4];
    long timestamp;
    char* path;
    char hash[9];
};   

// parse line and return an Event struct
struct Event parseLineIntoEvent(char* line) {
    struct Event event;
    char* lineSegment;

    int i = 0;
    lineSegment = strtok(line, " ");
    while (lineSegment != NULL) {
        if (i > 3) {
            printf("WARNING: input format error!\n");
            break;
        }
        if (i == 0)
            strncpy(event.action, lineSegment, sizeof(event.action)-1);
        else if(i == 1)
            event.timestamp = atoi(lineSegment);
        else if(i == 2) {
            event.path = malloc(sizeof(lineSegment));
            strcpy(event.path, lineSegment);
        } else if(i == 3)
            strncpy(event.hash, lineSegment, sizeof(event.hash)-1);
        lineSegment = strtok(NULL, " ");
        i++;
    } // while
    return event;
} // parseLineIntoEvent()

int main (int argc, const char * argv[]) {
//...
    printf("%s\n",line); //prints original line that was read in from stdin
    struct Event event = parseLineIntoEvent(line);
    printf("event action: %s\n", event.action);
    printf("event timestamp: %lu\n", event.timestamp);
    printf("event path: %s\n", event.path);
    printf("event hash: %s\n", event.hash);
    free(event.path);
    free(line);
//...
    return 0;
}

EDIT:
I read in a line with this function, which gets rid of the newline character:

// read in line from stdin, eliminating newline character if present
char* getLineFromStdin() {
    char *text;
    int textSize = 50*sizeof(char);
    text = malloc(textSize);

    if ( fgets(text, textSize, stdin) != NULL ) {
        char *newline = strchr(text, '\n'); // search for newline character
        if ( newline != NULL ) {
            *newline = '\0'; // overwrite trailing newline
        }
    }
    return text;
}

Thanks in advance!

Answer 1

but I am perplexed by the fact that the other char array event.hash does not exhibit this type of behavior

You got unlucky. hash[8] may have gotten a ‘\0’ by sheer (bad-)luck.

Try setting it to something “random” before your strtok loop

    int i = 0;
    event.hash[8] = '_';             /* forcing good-luck */
    lineSegment = strtok(line, " ");
    while (lineSegment != NULL) {