A clean and robust way to parse URL strings in Objective C - ios

Question

I have a requirement to take a string that represents a URL that can be in many formats and standardise it so it conforms with the URL spec.
If the URL string does not have a scheme, or it has a scheme that is not 'http' or 'https', it should use a default scheme.
I wanted to use NSURLComponents but if a scheme is not provided it parses the host as a path
NSURLComponents *components = [NSURLComponents componentsWithString:#"www.google.com.au"];
components.scheme = #"http";
NSLog(#"1: %#", components.path);
NSLog(#"2: %#", components.host);
NSLog(#"3: %#", components.string);
testtest[2619:869020] 1: www.google.com.au
testtest[2619:869020] 2: ((null))
testtest[2619:869020] 3: http:www.google.com.au <-- Invalid
Therefore I ended up with this category on NSString
#define DEFAULT_SCHEME #"http"
#implementation NSString (standardiseUrlFormat)
- (NSString*)standardiseUrlFormat {
NSURLComponents *components = [NSURLComponents componentsWithString:self];
BOOL hasScheme = components.scheme != nil;
// If no scheme or an invalid scheme is provided, default to http
if (!hasScheme) {
// We have to use string concatenation here because NSURLComponents will
// put the hostname as the path if there is no scheme
return [NSString stringWithFormat:#"%#://%#", DEFAULT_SCHEME, self];
}
// Now we know that a scheme exists, check if it is a correct scheme
if (![components.scheme isEqualToString:#"http"] &&
![components.scheme isEqualToString:#"https"]) {
// Overwrite scheme if not supported
components.scheme = DEFAULT_SCHEME;
}
return [components string];
}
#end
With the following output
NSLog(#"1: %#", [#"http://www.google.com" standardiseUrlFormat]);
NSLog(#"2: %#", [#"www.google.com" standardiseUrlFormat]);
NSLog(#"3: %#", [#"https://www.google.com" standardiseUrlFormat]);
NSLog(#"4: %#", [#"https://www.google.com/some_path" standardiseUrlFormat]);
NSLog(#"5: %#", [#"www.google.com/some_path" standardiseUrlFormat]);
testtest[7411:944022] 1: http://www.google.com
testtest[7411:944022] 2: http://www.google.com
testtest[7411:944022] 3: https://www.google.com
testtest[7411:944022] 4: https://www.google.com/some_path
testtest[7411:944022] 5: http://www.google.com/some_path
Can anyone suggest a cleaner solution that doesn't use two methods (NSURLComponents and string concatenation) to construct the string?

Answer 1

Don't use string concatenation at all. Use NSURLComponents to form the desired NSURL; that's what it's for. For example, if you don't like what the scheme is, set the scheme to what you do want.
EDIT I guess I was thinking that having detected that this is a hostless URL you would rejigger it by hand, e.g.
let s = "www.apple.com/whatever" as NSString
let arr = s.pathComponents
let c = NSURLComponents()
c.scheme = "http"
c.host = arr[0]
c.path = "/" + (Array(arr.dropFirst()) as NSArray).componentsJoinedByString("/")
But perhaps this can't be done, and the problem really is that a URL without a scheme is more or less not a URL.