Question

Is there some standard method to remove all characters from string that cannot be used for a filenames in osx? Does the characters you are allowed to use correspond to some commonly used category? (I know defining a list of allowed characters would be possible, but I want to be able to use filenames from several languages (chinese, among others), so this will be a bit of a pain.) A solution that works in ubuntu too, would be great.

Answer 1

I use the following code to convert a arbitrary string into a valid filename. It follows similar rules (if not the same) to those applied to creating a webloc file when dragging a URL to the Desktop.

// Returns a sting representing this filename transformed (if neccesary) to make
// a valid (Mac OS X) filename.
- (NSString *) stringByMakingFileNameValid:(NSString *)fileName
{
    NSMutableString * validFileName = [NSMutableString stringWithString:fileName];
    if (!validFileName || [validFileName isEqualToString:@""]) {
        return @"untitled";
    }
    // remove initial period chars "."
    if ([validFileName hasPrefix:@"."]) {
        NSRange dotRange = [validFileName rangeOfCharacterFromSet:[NSCharacterSet characterSetWithCharactersInString:@"."]];
        [validFileName deleteCharactersInRange:dotRange];
    }
    // remove any colon chars ":" (same as webloc creation behaviour)
    [validFileName replaceOccurrencesOfString:@":" withString:@"" options:0 range:NSMakeRange(0, [validFileName length])];
    // this may lead to spaces at either end which need trimming
    validFileName = [[[validFileName stringByTrimmingCharactersInSet:[NSCharacterSet whitespaceCharacterSet]] mutableCopy] autorelease];

    // if there is nothing left return default value
    if ([validFileName isEqualToString:@""]) {
        return @"untitled";
    }
    // replace other disallowed Mac OS X characters
    [validFileName replaceOccurrencesOfString:@"/" withString:@"-" options:0 range:NSMakeRange(0, [validFileName length])];

    // if grater than 102 chars reduce to 101 and add elipses
    if ([validFileName length] > 102) {
        [validFileName deleteCharactersInRange:NSMakeRange(100, [validFileName length]-100)];
        [validFileName appendString:@"…"];
    }

    return [[validFileName copy] autorelease];
}

At the Darwin level “/” characters are stored as “:” characters (and vice versa). They are removed here for cosmetic reasons.