help with a regular expression

[mayonaise]

i can't seem to figure out how to formulate this regular expression..

i want to have my parser replace all newline characters "\n" with HTML breaks "<br/>". but ONLY if the newline characters are not preceeded immediately by certain patterns. say for example, the '@' symbol:

"this is a test@\n
and i don't want a BR before this line"

is being transformed into this:

"this is a test@<br/>
and i don't want a BR before this line"

i do not want that <br/> to show up in thie case, because it is preceeded immediately by the '@' character. nevertheless, i can't figure out how to write a regular expression that will ignore the preceeding pattern if found.

make sense? is this even possible with PHP's regular expressions?

i'm using preg_replace() to invoke the regular expression replacement.

[Frederick]

If this is like an articles or large block of text, just use the nr2br() function.
http://us2.php.net/manual/en/function.nl2br.php

<?php echo nl2br($article_text); ?>

[xblue]

you may want to check out lookbehinds which the preg functions do support. syntax for a negative lookbehind
[pre]
(?<!...)
[/pre]

[Installer]

Here's a generic solution for your problem, and ones similar to it:

// function str_replace_conditional($str, $arr, $srch, $repl, $delete = true)
// Replaces $srch with $repl in $str,
// unless $srch is immediately preceded by any string in $arr.
// If $delete == true, $srch preceded by a value in $arr will be deleted,
// else $srch will be left in place.
// All strings may be any length.
// Returns processed $str.
function str_replace_conditional($str, $arr, $srch, $repl, $delete = true)
{
    $offset = 0;
    while (($pos = strpos($str, $srch, $offset)) !== false) {
        $replaced = false;
        foreach ($arr as $each) {
            $len = strlen($each);
            if (substr($str, $pos - $len, $len) == $each) {
                if ($delete) {
                    $str = substr_replace($str, '', $pos, strlen($srch));
                }
                $replaced = true;
                break;
            }
        }
        if (!$replaced) {
            $str = substr_replace($str, $repl, $pos, strlen($srch));
        }
        $offset = $pos + 1;
    }
    return $str;
}

Tested, but not strenuously.
Not meant to compete with any regular expression solution.

Demo script:

$string = "\nadmin@\nexample.com \n mail@example.\ncom &\nhello\n";
$array = array('@', '&');
$search = "\n";
$replace = '<br />';

$new_string = str_replace_conditional($string, $array, $search, $replace);

echo '<pre>' . htmlentities($string) . '</pre><br />';
echo $string . '<br />';
echo '____________________' . '<br />';
echo '<pre>' . htmlentities($new_string) . '</pre><br />';
echo $new_string . '<br />';

Output:

admin@
example.com 
 mail@example.
com &
hello

admin@ example.com mail@example. com & hello 
____________________

<br />[e-mail] <br /> mail@example.<br />com &hello<br />

[email][e-mail][/email] 
mail@example.
com &hello