[SOLVED] Calling a User Function Within a Regex

[Push Eject]

Hi all.  I was unsure whether to post this in the Regex forum or here.  I hope I've chosen correctly.

 

I'm trying to call a routine to obfuscate email addresses from within a regular expression and am stumped.

 

The following code:

$body = a post to a page that may include an email address.

$body = eregi_replace('([_\.0-9a-z-]+@([0-9a-z][0-9a-z-]+\.)+[a-z]{2,3})', encodeEmail('\\1'), $body);

function encodeEmail($email){
$t = "mailto:" . $email;
$r = "";

for( $i = 0; $i < strlen($t); $i++ ){
	$r .= "&#" . ord( substr($t,$i,1) );
}

$r = "<a href=\"" . $r . "\">".$r."</a>";

return $r;
}

 

Returns:

<a href="&#109&#97&#105&#108&#116&#111&#58&#49">&#109&#97&#105&#108&#116&#111&#58&#49</a>
OR
mailto:1

 

Can I not call a user function from within the regex?  Am I referencing "\\1" wrong?

 

More confusing to me is that if I return $email from encodeEmail() it passes the correct string (\\1) back.

 

Any help is greatly appreciated.

 

Thanks,

Charlie

[Push Eject]

bump

[Push Eject]

Sorry I didn't format the code for color.  I'm such a n00b.   Here it is again in a more readable state:

 

<?php
$body = "blah blah john@example.com blah blah".

$body = eregi_replace('([_\.0-9a-z-]+@([0-9a-z][0-9a-z-]+\.)+[a-z]{2,3})', encodeEmail('\\1'), $body);

function encodeEmail($email){
$t = "mailto:" . $email;
$r = "";

for( $i = 0; $i < strlen($t); $i++ ){
	$r .= "&" . "#" . ord( substr($t,$i,1) );
}

$r = "<a href=\"" . $r . "\">".$r."</a>";

return $r;
}
?>

 

Returns:

blah blah <a href="&#109&#97&#105&#108&#116&#111&#58&#49">&#109&#97&#105&#108&#116&#111&#58&#49</a> blah blah

OR

mailto:1

 

Been playing with this again this morning and cannot figure out how to properly pass the "\\1" value...

[discomatt]

http://php.net/preg_replace_callback

[Push Eject]

Thanks Disco!

 

I *just* found that function myself and here is my solution:

<?php
$body = preg_replace_callback('([_\.0-9a-z-]+@([0-9a-z][0-9a-z-]+\.)+[a-z]{2,3})', 'encodeEmail', $body);

function encodeEmail($email){

$t = $email[0];
$r = "";

for( $i = 0; $i < strlen($t); $i++ ){
	$r .= "&" . "#" . ord( substr($t,$i,1) );
}

$r = "<a href=\"&#109a&#105l&#116o&#58" . $r . "\">".$r."</a>";

return $r;
}
?>

[discomatt]

Just for efficiency's sake.. you shouldn't need a capturing group in your regex, as index 0 will contain the entire match already.

[thebadbad]

If you wan't to do it similar to what you tried, the /e modifier treats the replacement as PHP code:

 

<?php
$body = "blah blah john@example.com blah blah".

$body = preg_replace('~([-_.0-9a-z]+@([0-9a-z][-0-9a-z]+\.)+[a-z]{2,3})~ei', "encodeEmail('\\1')", $body);

function encodeEmail($email){
$t = "mailto:" . $email;
$r = "";

for( $i = 0; $i < strlen($t); $i++ ){
	$r .= "&" . "#" . ord( substr($t,$i,1) );
}

$r = "<a href=\"" . $r . "\">".$r."</a>";

return $r;
}
?>

 

Also, in regular expressions you shouldn't escape a dot in a character class, so I removed the backslash (don't know if it allows a backslash in your code - we don't want it to). And I moved hyphens to the front of the classes, else it's defining a range (when not escaped). At last I made the search case insensitive by adding the /i modifier.

 

Also note that TLD's (the 'extension' of a domain) can be longer than 3 chars.

[Push Eject]

Thanks, everybody, this is great!

[discomatt]

If you wan't to do it similar to what you tried, the /e modifier treats the replacement as PHP code:

Also, in regular expressions you shouldn't escape a dot in a character class, so I removed the backslash (don't know if it allows a backslash in your code - we don't want it to). And I moved hyphens to the front of the classes, else it's defining a range (when not escaped). At last I made the search case insensitive by adding the /i modifier.

 

Also note that TLD's (the 'extension' of a domain) can be longer than 3 chars.

 

preg_replace_callback is a nicer than using the 'e' modifier IMO. Something about eval'ing generated data scares me

Hyphens can also be at the end of a character class. [-a-z] == [a-z-]

[thebadbad]

preg_replace_callback is a nicer than using the 'e' modifier IMO. Something about eval'ing generated data scares me

Hyphens can also be at the end of a character class. [-a-z] == [a-z-]

 

I think it's easier to control what goes where (concerning function parameters) when using the /e modifier, but I see what you mean. And yeah, you're right about the hyphens

[discomatt]

The 'e' modifier is nice for plugging data into an existing function... but I'd still rather create/use a handler...

 

<?php

function exiting_function ( $arg1, $arg2, $arg3 ) {
# Do something!?
}

function existing_function_handler ( $matches ) {
return existing_function( $matches[1], $matches[2], $matches[3] );
}

?>

 

Saves you from having to open up a potential hole