need help understanding this function

Hi all I'm trying to figure out how this function works, pointers give me a hard time still especially a function with several pointers as parameters.

I understand the function takes a string and if it has a % or + it removes those, I don't understand the nibble part of the function is where I'm confused. Long story short guess I just wonder if anyone could comment it so I can understand what's happening.

Code:

/* Unencode a urlencoded string*/
/*----------------------------------------------------------------------------*/
static void unencode(char *dst,char *src,size_t len)
{	char c;
	while (c = *src++,len--)
	{	if (c == '+') c = ' ';
		else if (c == '%')
		{	c = *src++;
			c = (nibble(c) << 4) | nibble(*src++);
			len -= 2;
		}
		*dst++ = c;
	}
	*dst = '\0';
}

Well a nibble is just 4 bits of a byte, and can be represented as a single character 0-9A-F.

The % in URL encoding is followed by 2 hex digits, say %41 would be a long winded way of saying 'A'

Doing a bit of research on urlencodeing and decoding you have a string with a single char being represented by a % and a 2 digit hex code for the actual char in string format so "%22" would be equal to " and "%28" would equal to (

Since the hex digits are in string format you have to convert the ascii char '2' into the actual hex number. A nibble is just a single hex digit. A high and low nibble represent a single char.

So it is just a string to hex conversion.


[edit]
That isn't very portable is it? I mean wouldn't there be endian issues?

Originally Posted by manofsteel972

Doing a bit of research on urlencodeing and decoding you have a string with a single char being represented by a % and a 2 digit hex code for the actual char in string format so "%22" would be equal to " and "%28" would equal to (

Since the hex digits are in string format you have to convert the ascii char '2' into the actual hex number. A nibble is just a single hex digit. A high and low nibble represent a single char.

So it is just a string to hex conversion.


[edit]
That isn't very portable is it? I mean wouldn't there be endian issues?

Endian issues only arise if you have "words" longer than a byte. Since all (that I know of) encodings used for URL's are single-byte, it shouldn't be a problem - I know of no processor that stores NIBBLES "the wrong way around".

--
Mats

Ok. So a single byte char is the only data type not affected by endian issues. And wide char or multibyte unicode isn't affected either since it is still handled as bytes right?

[edit] Answered my own question. A wide char is affected by endian issues.