如何替换不在 JavaScript 中特定标签内的文本

Question

I have a string (partly HTML) where I want to replace the string :-) into bbcode :wink: . But this replacement should not happen within <pre> , but in any other tag (or even not within a tag).

For example, I want to replace

:-)<pre>:-)</pre><blockquote>:-)</blockquote>

to:

:wink:<pre>:-)</pre><blockquote>:wink:</blockquote>

I already tried it with the following RegEx, but it does not work (nothing gets replaced):

var s = ':-)<pre>:-)</pre><blockquote>:-)</blockquote>';
var regex = /:\-\)(?!(^<pre>).*<\/pre>)/g;
var r = s.replace(regex, ':wink:');

Can someone please help me? :-)

Answer 1

You could avoid hellish regexes altogether if you use a suitable library such as jQuery, eg:

var excludeThese = ['pre'];

// loop over all elements on page, replacing :-) with :wink: for anything
// that is *not* a tag name in the excludeThese array

$('* not:(' + excludeThese.join(',') + ')').each(function() {
    $(this).html($(this).html().replace(/:\-\)/,':wink:'));
});

Answer 2

This ought to do it:-

var src = ":-)<pre>:-)</pre><blockquote>:-)</blockquote>"

var result = src.replace(/(<pre>(?:[^<](?!\/pre))*<\/pre>)|(\:\-\))/gi, fnCallback)

function fnCallback(s)
{
    if (s == ":-)") return ":wink:"
    return s;
}

alert(result);

It works because any pre element will get picked up by the first option in the regex and once consumed means that any contained :-) can't be matched since the processor will have moved beyond it.

Answer 3

Just thought it'd be worth offering a DOM solution:

Eg

var div = document.createElement('div');
div.innerHTML = ":-)<pre>:-)</pre><blockquote>:-)</blockquote>";

replace(div, /:-\)/g, ":wink:", function(){

    // Custom filter function.
    // Returns false for <pre> elements.

    return this.nodeName.toLowerCase() !== 'pre';

});

div.innerHTML; // <== here's your new string!

And here's the replace function:

function replace(element, regex, replacement, filter) {

    var cur = element.firstChild;

    if (cur) do {

        if ( !filter || filter.call(cur) ) {

            if ( cur.nodeType == 1 ) {
                replace( cur, regex, replacement );
            } else {
                cur.data = cur.data.replace( regex, replacement );
            }

        }

    } while ( cur = cur.nextSibling );

}

Answer 4

尝试使用var regex = /：-）（？！（^）* </ pre>）/ g;

Answer 5

Almost good: Your negative lookbehind and lookahead where not in the right position and need a slight adjustment:

/(?<!(<pre>)):-\)(?!(<\/pre>))/g

1. Looks for all ":-)"
2. ...but not if there is a <pre> behind (the regex cursor is!)
3. ...but not if there is a </pre> before (the regex cursor is!)

as a side effect though: <pre>:-):-)</pre> works too but not <pre>:-):-):-)</pre>

https://regex101.com/r/CO0DAD/1

ps. this is a firefox 104 browser (could be different in others)