Ah, yes. It seems Java doesn't do negative groups. The idea behind the [^(*/)]* block is to ensure that the comment block stops at the first *\, not some secondary comment close. For example, /* * / * */ width: 0; /**/ should turn into width:0;.
Nice to know about \s, I tend to overload the characters after dealing with quirky RegEx engines. I'll update the gist.
You made me think about the problem differently, which led to http://stackoverflow.com/questions/2503413/regular-expression-stop-at-first-match
I changed the RegEx block in question to .*?. This seems to pass my all of my tests so far.
Hmmm, I think you're a bit confused - Java does do negative character classes ([^abc] matches anything except a or b or c), and negative lookaheads ((?!abc) - matches any position not followed by abc string) - the latter is what you were attempting to do (with mixed up syntax), and would be /\*(?:(?!\*/).)*+\*/ if you were going to do it that way, but yeah - the simpler solution in this case is a lazy quantifier.
Have you tested against multi-line comments? By default . wont match a newline, so you may need to enable "dotall" mode, which can be done by placing a (?s) at the start of the pattern.
(The s is because other regex implementations call it "single-line" mode, just to be confusing.)
Java does negative characters, but it doesn't do negative groups. [^(abc)] should be anything except the string "abc."
Thanks again, the multiline test was failing without (?s). I had it buried where it wasn't easily seen. I changed my tests to be more visual and added them to the gist.
Still not sure where "negative groups" is coming from - I've not heard that term, nor seen the syntax [^(abc)] except as a negative character class (any char except those five) - not in Java/Python/Perl/PCRE or any others I'm aware of. Have you got a reference for it?
I don't have a reference offhand. Perhaps it's something floating in my brain from formal languages. It'd be rather handy. I know this isn't the first time I've tried to use it.
You don't need to write
[\s\t\r\n]because\sis shorthand for[ \t\r\n](also includes a few other whitespace characters you don't care about).You also don't need to escape
/nor do you need to escape*inside the character class, and not sure why you've got parentheses in there? :SActually, I'd probably just recommend using something like the YUI compressor, since that's built for this task and has already been extensively tested and so on.