在正则表达式上前瞻

Question

I am looking to have all my bbcode images that have a link attached to them be parsed before the normal [img] tag parsing is done in my code, this is because my normal [img] tags if they have no link get turned into a lightbox.

    $find = array(
    "/\[url=(.+?)\](?=\[img\](.+?)\[\/img\])\[\/url\]/is"
    );

    $replace = array(
    "<a href=\"$1\" target=\"_blank\"><img src=\"$2\" class=\"bbcodeimage img-polaroid\" alt=\"[img]\" /></a>"
    );

    $body = preg_replace($find, $replace, $body);

Totally wrong since it doesn't actually do anything it seems I am completely confused on the whole lookahead thing to capture [img] tags that have an [url] before it and replace accordingly.

Answer 1

I my opinion you don't need lookahead at all. Try this:

$find = array(
    '~\[url=([^]]+)]\[img]([^[]+)\[/img]\[/url]~i'
);

$replace = array(
    '<a href="$1" target="_blank"><img src="$2" class="bbcodeimage img-polaroid" alt="[img]" /></a>'
);

Explanations:

First at all, I have changed the pattern delimiter to ~, the goal of this change is to avoid to escape all literal / in the pattern. Literals ] don't need to be escaped outside a character class or inside a character class if (and only if) it is the first character.

A lookahead is not useful in this situation because a lookahead is only a check and matches nothing. Example a(?=bc) will find a a followed by bc but will only match the a. It is why lookaheads and lookbehinds are also called "zero width assertions".

pattern details:

~           # delimiter
\[url=      # literal: [url=
(           # open the first capturing group
    [^]]+   # all characters except ] (one or more times)
)           # close the first capturing group
]           # literal: ]
\[img]      # literal: [img]
(           # open the second capturing group
    [^[]+   # all characters except [ (one or more times)
)           # close the second capturing group
\[/img]     # literal: [/img]
\[/url]     # literal: [/url]
~i          # delimiter and i modifier

Note that I have choosen to use single quotes for the replacement string to avoid to escape all double quotes of the string (and because there is no reason to use double quotes, no variables, no or \t etc.).