regex-email.php

<?php 

// This code/file/output has no license, though attribution would be appreciated. 
// Output at http://pastebin.com/2ZLehM5N

if ($_SERVER["QUERY_STRING"] == "source") { 
    highlight_file(__FILE__); 
    return; 
} 

$ALPHA  = "[A-Za-z]"; 
$CR     = "\\r"; 
$CRLF   = "\\r\\n"; 
$DIGIT  = "[0-9]"; 
$DQUOTE = "\\x22"; 
$LF     = "\\n"; 
$VCHAR  = "[\\x21-\\x7E]"; 
$WSP    = "[\\x20\\t]"; 

/* 
 * RFC 5322: Internet Message Format <http://tools.ietf.org/html/rfc5322> 
 *  
 * $ADDR_SPEC is the form of an email address 
 *  
 * There is special PCRE syntax used to deal with the $COMMENT/$CCONTENT recursion, namely 
 * using (?P<foo>...) to name the first instance of the subpattern and (?P>foo) to recurse. 
 * Since there can't be multiple subpatterns with the same name, later instances of the subpattern 
 * use (?P>foo) to reference it. 
 *  
 * Note: this will validate an email address, but the SMTP standards have their own additional rules, 
 * such as fully-qualified domain names and maximum lengths. Just because this considers an email 
 * to be valid, SMTP may consider it invalid. Additionally, SMTP agents may take shortcuts and 
 * falsely accept or reject addresses when they should not. 
 */ 

// 4.1 Miscellaneous Obsolete Tokens 
$OBS_NO_WS_CTL = "[\\x01-\\x08\\x0B\\x0C\\x0E-\\x1F\\x7F]"; 
$OBS_CTEXT     = "$OBS_NO_WS_CTL"; 
$OBS_QP        = "\\\\(\\0| $OBS_NO_WS_CTL | $LF | $CR )"; 
$OBS_QTEXT     = "$OBS_NO_WS_CTL"; 

// 3.2.1 Quoted characters 
$QUOTED_PAIR = "(\\\\( $VCHAR | $WSP )| $OBS_QP )"; 

// 4.2 Obsolete Folding White Space 
$OBS_FWS = "( $WSP )+( $CRLF ( $WSP )+)*"; 

// 3.2.2 Folding White Space and Comments 
$FWS      = "((( $WSP )* $CRLF )?( $WSP )+| $OBS_FWS )"; 
$CTEXT    = "([\\x21-\\x27\\x2A-\\x5B\\x5D-\\x7E]| $OBS_CTEXT )"; 
$CCONTENT = "( $CTEXT | $QUOTED_PAIR | (?P>COMMENT) )"; 
$_COMMENT = "(?P<COMMENT>\\((( $FWS )? $CCONTENT )*( $FWS )?\\))"; 
$COMMENT  = '$COMMENT'; 
$CFWS     = "((( $FWS )? $COMMENT )+( $FWS )? | $FWS )"; 

// 3.2.3 Atom 
$ATEXT         = "( $ALPHA | $DIGIT |[!#\$%&'*+\\-/=?^_`{|}~])"; 
$ATOM          = "( $CFWS )?( $ATEXT )+( $CFWS )?"; 
$DOT_ATOM_TEXT = "( $ATEXT )+(\\.( $ATEXT )+)*"; 
$DOT_ATOM      = "( $CFWS )? $DOT_ATOM_TEXT ( $CFWS )?"; 

// 3.2.4 Quoted Strings 
$QTEXT         = "([\\x21\\x23-\\x5B\\x5D-\\x7E] | $OBS_QTEXT )"; 
$QCONTENT      = "( $QTEXT | $QUOTED_PAIR )"; 
$QUOTED_STRING = "( $CFWS )? $DQUOTE (( $FWS )? $QCONTENT )*( $FWS )? $DQUOTE ( $CFWS )?"; 

// 3.2.5 Miscellaneous Tokens 
$WORD = "( $ATOM | $QUOTED_STRING )"; 

// 4.4 Obsolete Addressing 
$OBS_LOCAL_PART = "$WORD (\\. $WORD )*"; 
$OBS_DOMAIN     = "$ATOM (\\. $ATOM )*"; 
$OBS_DTEXT      = "( $OBS_NO_WS_CTL | $QUOTED_PAIR )"; 

// 3.4.1 Addr-Spec Specification 
$LOCAL_PART     = "( $DOT_ATOM | $QUOTED_STRING | $OBS_LOCAL_PART )"; 
$DTEXT          = "([\\x21-\\x5A\\x5E-\\x7E]| $OBS_DTEXT )"; 
$DOMAIN_LITERAL = "( $CFWS )? \\[(( $FWS )? $DTEXT )*( $FWS )?\\]( $CFWS )?"; 
$DOMAIN         = "( $DOT_ATOM | $DOMAIN_LITERAL | $OBS_DOMAIN )"; 
$ADDR_SPEC      = "$LOCAL_PART @ $DOMAIN"; 

/// 

// deal with the recursion problem 
// replace the first $COMMENT with the correct expression, then subsequent ones with the recursive call 
$first = strpos($ADDR_SPEC, $COMMENT); 
$ADDR_SPEC = substr($ADDR_SPEC, 0, $first) . $_COMMENT . substr($ADDR_SPEC, $first + strlen($COMMENT)); 
$ADDR_SPEC = str_replace($COMMENT, "(?P>COMMENT)", $ADDR_SPEC); 

// remove whitespace, escape incidental slashes, add ^ and $ anchors 
$regex = '/^' . str_replace("/", "\\/", preg_replace('/\s+/', "", $ADDR_SPEC)) . '$/'; 

?> 
<pre style="word-wrap:break-word;"> 
<?= htmlentities($regex) ?> 
</pre> 
<a href="<?= basename(__FILE__) ?>?source">Source rules</a>