Question

I implemented the Pattern class as shown here:
http://www.java2s.com/Code/Java/GWT/ImplementjavautilregexPatternwithJavascriptRegExpobject.htm

And I would like to use the following regex to match urls in my String:

(http|https):\/\/(\w+:{0,1}\w*@)?(\S+)(:[0-9]+)?(\/|\/([\w#!:.?+=&%@!\-\/]))?

Unfortunately, the Java compiler of course fails on parsing that string because it doesn’t use valid escape sequences (since the above is technically a url pattern for JavaScript, not Java)

At the end of the day, I’m looking for a regex pattern that will both compile in Java and execute in JavaScript correctly.

Answer 1

You will have to use JSNI to do the regex evaluation part in Javascript. If you do write the regex with the escaped backslashes, that will get converted to Javascript as it is and will obviously be invalid. Thought it will work in the Hosted or Dev mode as thats still running Java bytecode, but not on the compiled application.

A simple JSNI example to test if a given string is a valid URL:

// Java method
public native boolean isValidUrl(String url) /*-{
    var pattern = /(http|https):\/\/(\w+:{0,1}\w*@)?(\S+)(:[0-9]+)?(\/|\/([\w#!:.?+=&%@!\-\/]))?/;
    return pattern.test(url);
}-*/;

There may be other irregularities between the Java and Javascript regex engines, so it’s better to offload it completely to Javascript at least for moderately complex regexes.