Simple string parsing in Cocoa / Objective-C: pars

Question

Here's a piece of code to take a string (either NSString or NSAttributedString) input that represents a command line and parse it into two strings, the command cmd and the arguments args:

NSString* cmd = [[input mutableCopy] autorelease];
NSString* args = [[input mutableCopy] autorelease];
NSScanner* scanner = [NSScanner scannerWithString:[input string]];
[scanner scanUpToCharactersFromSet:[NSCharacterSet 
                                    whitespaceAndNewlineCharacterSet] 
                        intoString:&cmd];
if (![scanner scanUpToString:@"magicstring666" intoString:&args]) args = @"";

That seems to work but the magic string is a pretty absurd hack. Also, I'm not at all sure I'm doing things right with the autoreleases.

ADDED: The solution should also be robust to initial whitespace. Also, I originally had the input string called both input and inStr. Sorry for that confusion.

ADDED: I believe one thing the above code gets right that the answers so far don't is that args should not have any initial whitespace.

Answer 1

NSString *cmd;
NSScanner *scanner = [NSScanner scannerWithString:[inStr string]];
[scanner scanUpToCharactersFromSet:[NSCharacterSet
                                    whitespaceAndNewlineCharacterSet] 
                        intoString:&cmd];
NSString *args = [[scanner string] substringFromIndex:[scanner scanLocation]];

Your autoreleases were OK, but allocating strings in the first place was unnecessary since NSScanner returns a new string by reference. Since NSScanner's charactersToBeSkipped include whitespace by default, it shouldn't get tripped up by initial whitespace.

Answer 2

If you want to expand the string into a full array of arguments like the input to 'main', you can use wordexp.

#import <wordexp.h>

+ (NSMutableArray*) splitArgumentString:(NSString*)strArgs
{
    wordexp_t expandedArgs;
    NSMutableArray *argArray = [NSMutableArray array];

    if(strArgs != nil && 0 == wordexp([strArgs UTF8String], &expandedArgs, 0))
    {
        for(size_t i = 0; i < expandedArgs.we_wordc; ++i)
        {
            NSString arg = [NSString stringWithCString:expandedArgs.we_wordv[i] encoding:NSUTF8StringEncoding];
            [argArray addObject:arg];
        }
        wordfree(&expandedArgs);
    }
    return argArray;        
}

Answer 3

The autoreleases you mentioned don't actually make any sense, all you're doing is creating a mutable copy (NSMutableString *) that's properly autoreleased, but since you're casting it to an NSString * there's no practical difference then just saying cmd = input;. Even that's not needed for args though, since NSScanner will overwrite what's there anyway.

The rangeOfString: would work, if you want to go this route you can trim leading whitespaces using NSString's stringByTrimmingCharactersInSet method (I would also test to be sure both arguments and the command exist, or you'll get an error). What I would do though, is use the NSString componentsSeparatedByCharactersInSet: method. This will give you an NSArray object containing the command and each argument in a separate index.

Answer 4

Something like this?

int index = [input rangeOfString:@" "].location;
NSString *cmd = [input substringToIndex:index]);
NSString *args = [input substringFromIndex:index + 1]);