Regex to strip BBCode

Question

I need a regular expression to strip out any BBCode in a string. I've got the following (and an array with tags):

new RegExp('\\[' + tags[index] + '](.*?)\\[/' + tags[index] + ']');

It picks up [tag]this[/tag] just fine, but fails when using [url=http://google.com]this[/url].

What do I need to change? Thanks a lot.

Answer 1

You have to allow any character other than ']' after a tag until you find ' ]'.

new RegExp('\\[' + tags[index] + '[^]]*](.*?)\\[/' + tags[index] + ']');

You could simplify this to the following expression.

\[[^]]*]([^[]*)\[\\[^]]*]

The problem with that is, that it will match [WrongTag]stuff[\WrongTag], too. Matching nested tags requires using the expression multiple times.

Answer 2

I think

new RegExp('\\[' + tags[index] + '(=[^\\]]+)?](.*?)\\[/' + tags[index] + ']');

should do it. Instead of group 1 you have to pick group 2 then.

Answer 3

To strip out any BBCode, use something like:

string alltags = tags.Join("|");
RegExp stripbb = new RegExp('\\[/?(' + alltags + ')[^]]*\\]');

Replace globally with the empty string. No extra loop necessary.

Answer 4

You can check for balanced tags using a backreference:

 new RegExp('\\[(' + tags.Join('|') + ')[^]]*](.*?)\\[/\\1]');

The real problem is that you cant't match arbitrary nested tags in a regular expression (that's the limit of a regular language). Some languages do allow for recursive regular expressions, but those are extensions (that technically make them non-regular, but doesn't change the name that most people use for the objects).

If you don't care about balanced tags, you can just strip out any tag you find:

 new RegExp('\\[/?(?:' + tags.Join('|') + ')[^]]*]');

Answer 5

I came across this thread and found it helpful to get me on the right track, but here's an ultimate one I spent two hours building (it's my first RegEx!) for JavaScript and tested to work very well for crazy nests and even incorrectly nested strings, it just works!:

string = string.replace(/\[\/?(?:b|i|u|url|quote|code|img|color|size)*?.*?\]/img, '');

If string = "[b][color=blue][url=www.google.com]Google[/url][/color][/b]" then the new string will be "Google". Amazing.

Hope someone finds that useful, this was a top match for 'JavaScript RegEx strip BBCode' in Google ;)

Answer 6

Remember that many (most?) regex flavours by default do not let the DOT meta character match line terminators. Causing a tag like

"[foo]dsdfs
fdsfsd[/foo]"

to fail. Either enable DOTALL by adding "(?s)" to your regex, or replace the DOT meta char in your regex by the character class [\S\s].