将URL转换为链接，除非它们在使用PHP的BBCode中

Question

I have text that supports BBCode and I use a custom PHP function to do the replacing instead of using PECL package or anything. It's working great to turn:

[url=http://www.google.com]Google[/url]

to

<a href="http://www.google.com" target="_blank">Google</a>

I also have a custom BBCode tag that turns the tag into an iframe:

[customtag]abc[/customtag]

to

<iframe src="http://example.com/abc"></iframe>

But now I have a requirement to change urls that are just typed normally to turn them into links. So if you type:

http://www.google.com

it will turn it into:

<a href="http://www.google.com" target="_blank">Google</a>

I've used this regex to do this:

$string = preg_replace('@(https?://([-\w\.]+[-\w])+(:\d+)?(/([\w/_\.#-]*(\?\S+)?[^\.\s])?)?)@', '<a href="$1" target="_blank">$1</a>', $string);

But it's also matching what is in bbcode. If I include it after I parse the bbcode then it changes things in the anchor. Also happens for images and the iframe.

So how I can I change plain urls but ignore ones that are within a bbcode tag or already in a html tag?

Answer 1

This may not be the "best" solution, but you can use a negative lookbehind ((?<!...)) to make sure that the URL isn't prefixed by ', ", or =. The obvious limitations are if someone writes something like:

Let's visit "https://google.com" on our computers. Or the link=https://google.com.

Anyways, the negative lookbehind would go at the very beginning of your expression and contain a character class: ["'=].

(?<!["'=])(https?://([-\w\.]+[-\w])+(:\d+)?(/([\w/_\.#-]*(\?\S+)?[^\.\s])?)?)

Demo