Regex Unclosed character class in Java; escaping doesn't solve issue

Question

I have this regex in JS flavor

^[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#[\]@!\$&'\(\)\*\+,;=.]+$

I tried it in JS and it worked for me (only want to match URL that does not start with the HTTP/HTTPS protocol): https://regex101.com/r/6y2Gnd/2

Now I want to use the same regex in my Java backend. At first I got the error

Unclosed character class

Upon reading into it, I realized I have to escape the \ slash. I basically added three \\\ to every \ slash. The result is:

^[\\\\w.-]+(?:\\\\.[\\\\w\\\\.-]+)+[\\\\w\\\\-\\\\._~:/?#[\\\\]@!\\\\$&'\\\\(\\\\)\\\\*\\\\+,;=.]+$

Even though the compiler doesn't show any errors anymore, the result was empty, i.e. it couldn't match the cases like it did with in JS flavor.

I tested the Java regex here and in my code.

www.web.de # I want to match this
web.de # I want to match this
http://web.de # I do NOT want to match this
https://www.web.de # I do NOT want to match this

Anyone know what I'm missing?

Why are you turning a single '\' into '\\\\'? Have you tried just '\\'? — Jason, Feb 23 '20 at 23:40
@Jason Yes, I've tried that too: `^[\\w.-]+(?:\\.[\\w\\.-]+)+[\\w\\-\\._~:/?#\[\\]@!\\$&'\$\$\\*\\+,;=.]+$`. Still no matches. I escaped because of Java. I've been told that java behaves like that. — thadeuszlay, Feb 23 '20 at 23:43
Like so? `^[\\w.-]+(?:\\.[\\w\\.-]+)+[\\w\\-\\._~:/?#\[\\]@!\\$&'\$\$\\*\\+,;=.]+$`? — thadeuszlay, Feb 23 '20 at 23:51

score 0 · Accepted Answer · answered Feb 23 '20 at 23:44

0

Following regex works well in Java regex tester:

^[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#\[\]@!\$&'\(\)\*\+,;=.]+$

Please try it yourself. Backslash should be added in the front of the [ character.

answered Feb 23 '20 at 23:44

Marko Medojević

595
5
8

I added `^[\w.-]+(?:\.[\w\.-]+)+[\w\-\._~:/?#\[\]@!\$&'\*\+,;=.]+$` into https://www.freeformatter.com/java-regex-tester.html#ad-output But no matches found – thadeuszlay Feb 23 '20 at 23:45
I just tested and it works. Multiline checkbox should be selected. – Marko Medojević Feb 23 '20 at 23:47
Yes, you are right! It's working – thadeuszlay Feb 23 '20 at 23:54

Regex Unclosed character class in Java; escaping doesn't solve issue

1 Answers1