I want to check if the characters of a string are in this form:
hw: + one numeric character + , + one numeric character
hw:0,0
hw:1,0
hw:1,1
hw:0,2
Et cetera
/* 'hw:' + 'one numeric character' + ',' + 'one numeric character' */
I found strncmp(arg, "hw:", 3) but that only checks the first 3 characters.
It's tempting to use strlen() and sscanf():
char *data = "hw:1,2";
char digit1[2];
char digit2[2];
if (strlen(data) == 6 && sscanf(data, "hw:%1[0-9],%1[0-9]", digit1, digit2) == 2)
...then the data is correctly formatted...
...and digit1[0] contains the first digit, and digit2[0] the second...
This works particularly well if you need to know what the two digits are, not merely that the format is correct. However, you can also pull the digits out by fixed position, so that is not crucial. It also upgrades gracefully (though not without change) if you need to allow for "hw:12,25" in the future.
strncmp(arg, "hw:", 3) is a good start (remember that the function returns zero when the match is found). Next, you need to check that
This results in the following expression:
if (!strncmp(arg, "hw:", 3) && isdigit(arg[3]) && arg[4] == ',' && isdigit(arg[5])) {
...
}
Note the use of isdigit(int) to test that a character is a digit.
If the numbers could span more than a single digit, you could use sscanf: this would also let you retrieve the values:
int a, b;
if (sscanf(arg, "hw:%d,%d", &a, &b) == 2) {
...
}
%d notations mean that hw: 3, 4 would be allowed as valid. So would hw:-34,+45. You can preclude those with a length check.The GNU C library supports regular expressions. If you don't want to learn regular expressions, you can just repeatedly use strncmp as well as functions from the ctype.h header.
