Question

Let’s say, I have a regular expression that checks the validation of the input value as a whole. For example, it is an email input box and when user hits enter, I check it against ^[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}$ to see if it is a valid email address.

What I want to achieve is, I want to intercept the character input too, and check every single input character to see if that character is also a valid character. I can do this by adding an extra regular expression, e.g. [A-Z0-9._%+-] but that is not what I want.

Is there a way to extract the widest possible range of acceptable characters from a given regular expression? So in the example above, can I extract all the valid characters that are defined by the original regular expression (i.e. ^[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,4}$) programmatically?

I would appreciate any help or hint.

P.S. This is project for iOS written in Objective-C.

Answer 1

If you don’t mind writing half a regex parser, certainly. You would have to be able to distinguish literals from meta-characters and to unroll/merge all character classes (including negated character classes, and nested negated character classes, if you regex flavor supports them).

If NSRegularExpressions doesn’t come with some convenience method, I cannot imagine how it would be possible otherwise. Just think about ^. When it is outside of a character class, it’s a meta-character that you can ignore. If it is inside a character class, it’s a meta-character, that negates the character class unless it is not the first character. - is a meta-character inside character classes, unless it is the first character, the last character, or right after another character range (depending on regex flavor). And I’m not even speaking about escaped characters.

I don’t know about NSRegularExpressions, but some flavors also support nested character classes (like [a-z[^aeiou]] for all consonants). I think you get where I am going with this.