Help with a Regular Expression

Lazesharp

Ive been trying to do this for ages now, i know its possible, but im unable to find the solution.

What regular expression would i need to change the string
[link=http://www.google.com]Google[/link]

into its html equivilent of
<a href="http://www.google.com>Google</a>

Any ideas?

Weedpacket

use preg_replace() (I say this in case you were thinking of using ereg_replace()...)

The search string is:

"#\[link=([^\]]+)\]((?:(?!\[/link\]).)*)[/link]#i"

And the replacement text will be

"<a href=\"\\1\">\\2</a>"

The critical parts of the search string (that get remembered in the replacmeent) are
[^]]+
(as many characters as possible after the =, up to but not including the ])
and
b*[/b]
Which matches zero or more characters unless they form the start of [/link]. That's the bit that would make it such a bitch to do with an ereg_replace.

Of course, there's the possibility of whitespace being introduced here and there in the [link=...]...[/link] pseudotags. Stick \s in judicious spots to avoid this, e.g., around the =:

"#\[link\s*=\s*([^\]]+)\]((?:(?!\[/link\]).)*)[/link]#i"

Note that \s* lie outside the () around the bits that are remembered - we're not interested in remembering if there was any whitespace there or not.

xblue

(oops, too slow again)

I wonder where you get stuck with this? Anyway, let's do it step by step:

1st part: [link= would be matched by [link=, you only need to escape the square bracket.

2nd: this is the url that may very. I don't think you want to validate the url in any way, so we can simple say [^] ]+ (which means a sequence of characters that are not the closing bracket)

3rd: ] (again, don't forget to escape)

4th: as 2nd, but limited by the next opening bracket [^[]+

5th: [/link] (see 1 and 3)

Now, put it all together, and put parts you need to re-use in parenthesis, and don't forget to escape the sign you use as delimiter:

/[link=([^{] ]+)]([^{[]+)[\/link] /}} now, that's our pattern - starts looking nasty, eh 😉?

okay, and the replacement argument. Here we can use $1 and $2 to get what was matched by the subpattern in the 1st and 2nd (), and the i-modifier to match case-insensitive:

preg_replace("/[link=([^] ]+)]([^[]+)[\/link] /i", "<a href=\"$1\">$2</a>", $string);

Hope there are no typos, but once you get into it you should be able to modify it according to your needs anyway.

Weedpacket

Originally posted by xblue
preg_replace("/[link=([^{] ]+)]([^[]+)[\/link] /i", "<a href=\"$1\">$2</a>", $string);}

The drawback to this expression is that it won't recognise a match in, e.g.

[link=http://www.google.com]A [big]very[/big] useful link.[/link]

xblue

Yes Weedpacket, I have to admit that, so your solution with lookaheads definitely catches more.

Another approach might be to use ungreedy matching:
preg_replace("/[link=(.+?)](.+?)[\/link] /si", "<a href=\"$1\">$2</a>", $string);

(I'm not sure how that compares to the other one with respect to efficiency, probably it's worse, but I think this should catch nested tags as well.)

Weedpacket

Originally posted by xblue
Another approach might be to use ungreedy matching:
preg_replace("/[link=(.+?)](.+?)[\/link] /si", "<a href=\"$1\">$2</a>", $string);

Y'know, I actually managed to forget about ungreedy matching! I agree that it would probably be less efficient (certainly for single characters), but it does work. And is a lot simpler.

Still be a bitch to do in ereg_replace, though... 🙂

Lazesharp

sorry about this, i tried all the methods here, and only the first one (from Weed) appeared to work.

Unfortunately [link=http://www.google.com]Google[/link] produces this:

<a href="http://www.google.com">Goog</a>e[/link]

Is there something im doing wrong?

xblue

well, you didn't like my suggestions 🙁, that's okay, I didn't test them, just tried to give you an idea how regexp work.

Anyway, I tried to explain this: if you use [] without escaping it, it is misunderstood.

That [/link] at the end of the expression means: 1 character out of any listed between the square brackets, so it must match the "l" in google. Escape the square brackets with a backslash:

[/link]

Lazesharp

Thanks a ton to both of you, xblue i did try your idea, its just the one i used seemed to work best, thanks anyway.

And the solution xblue posted works a charm, thx mate.

Lotta

I’ve got a question about those \s* .
I use this preg_replace to make block quotes in my forum

$message = preg_replace("/[QUOTE](.*)[\/QUOTE]/miU","<blockquote>\1</blockquote>",$ message, 1);

but it only works if the text is written like this:

text text [ QUOTE ] text text text [ /QUOTE ] text text

If the message has a line break after the [ /QUOTE ] like this:

text text [ QUOTE ] text text text [ /QUOTE ]
text text text text

the preg_replace will not work. How can I fix this?

Weedpacket

You need another modifier; miUs; without it, . won't match a newline.