CODEWITHSUNDEEP
iosobjective-cregex
Idan Moshe
Idan Moshe

Reputation: 1533

Checking for a valid Hebrew regex return always YES

I've a certain regex pattern to check against. Valid result is is only Hebrew language, letters, marks etc.

//////////Regex//////////
static NSString *const HEBREW_NUMBERS_NON_NUMERIC_CHAR = @"([\u0590-\u05FF]*|[0-9]*|[\\s]*|[.-:;,?!/&*()+=_'\"]*)+";


+ (BOOL)hasValidOpenLine:(NSString *)openLine
{
    if (openLine.length >= MIN_NUMBER_OF_CHARACTERS_IN_OPEN_LINE || openLine.length <= MAX_NUMBER_OF_CHARACTERS_IN_OPEN_LINE) {
        NSError *errorRegex;
        NSRegularExpression *regexOpenLine = [[NSRegularExpression alloc] initWithPattern:HEBREW_NUMBERS_NON_NUMERIC_CHAR
                                                                                  options:0
                                                                                    error:&errorRegex];
        NSRange range = NSMakeRange(0, openLine.length);

        if ([regexOpenLine numberOfMatchesInString:openLine options:0 range:range] > 0) {
            return YES;
        }
    }
    return NO;
}

But no matter what I type, it always return me YES even for only English string.

Upvotes: 0

Views: 200

Answers (1)

Itai Ferber
Itai Ferber

Reputation: 29918

There may be two things going wrong here, depending on your test string. First off, the stars in your regex allow for empty matches against strings which would otherwise not match, which is why your regex might match English strings — matching your regex on @"Hello, world!" returns {0, 0}, a range whose location is not NSNotFound, but whose length is zero.

The other issue is that you're not anchoring your search. This will allow the regex to match against singular characters in strings that would otherwise not match (e.g. the , in @"Hello, world!"). What you need to do is anchor the regex so that the whole string has to match, or else the regex rejects it.

Your modified code can look something like this:

static NSString *const HEBREW_NUMBERS_NON_NUMERIC_CHAR = @"([\u0590-\u05FF]|[0-9]|[\\s]|[.-:;,?!/&*()+=_'\"])+";

+ (BOOL)hasValidOpenLine:(NSString *)openLine
{
    if (openLine.length >= MIN_NUMBER_OF_CHARACTERS_IN_OPEN_LINE || openLine.length <= MAX_NUMBER_OF_CHARACTERS_IN_OPEN_LINE) {
        NSError *errorRegex;
        NSRegularExpression *regexOpenLine = [[NSRegularExpression alloc] initWithPattern:HEBREW_NUMBERS_NON_NUMERIC_CHAR
                                                                                  options:0
                                                                                    error:&errorRegex];
        if ([regexOpenLine numberOfMatchesInString:openLine options:NSMatchingAnchored range:NSMakeRange(0, openLine.length)] > 0) {
            return YES;
        }
    }

    return NO;
}

This will now match against strings like @"שלום!", and not strings like @"Hello, world!" or @"Hello: היי", which is what I assume you're going for.

In the future, if you're looking to debug regexes, use -[NSRegularExpression rangeOfFirstMatchInString:options:range:] or -[NSRegularExpression enumerateMatchesInString:options:range:usingBlock:]; they can help you find matches that may cause your regex to accept unnecessarily.

Upvotes: 2

Related Questions