PHP: imap_8bit

imap_8bit

umu
28-Jan-2006 05:58


...deed.ztinmehc-ut.zrh@umuumu@hrz.tu-chemnitz.deed...



I use the following function instead of imap_8bit 

when using PHP without the IMAP module, 

which is based on code found in 

quoted_printable_decode,

and giving (supposedly) exactly the same results as imap_8bit,

(tested on thousands of random strings containing lots

of spaces, tabs, crlf, lfcr, lf, cr and so on,

no counterexample found SO FAR:)



AND you can force a trailing space to be encoded,

as opposed to what imap_8bit does,

which I consider is a violation of RFC2045,

(see http://bugs.php.net/bug.php?id=35290) 

by commenting that one central line.



<?php

function quoted_printable_encode($sText,$bEmulate_imap_8bit=true) {

  // split text into lines

  $aLines=explode(chr(13).chr(10),$sText);



  for ($i=0;$i<count($aLines);$i++) {

    $sLine =& $aLines[$i];

    if (strlen($sLine)===0) continue; // do nothing, if empty



    $sRegExp = '/[^\x09\x20\x21-\x3C\x3E-\x7E]/e';



    // imap_8bit encodes x09 everywhere, not only at lineends,

    // for EBCDIC safeness encode !"#$@[\]^`{|}~,

    // for complete safeness encode every character :)

    if ($bEmulate_imap_8bit)

      $sRegExp = '/[^\x20\x21-\x3C\x3E-\x7E]/e';



    $sReplmt = 'sprintf( "=%02X", ord ( "$0" ) ) ;';

    $sLine = preg_replace( $sRegExp, $sReplmt, $sLine );  



    // encode x09,x20 at lineends

    {

      $iLength = strlen($sLine);

      $iLastChar = ord($sLine{$iLength-1});



      //              !!!!!!!!    

      // imap_8_bit does not encode x20 at the very end of a text,

      // here is, where I don't agree with imap_8_bit, 

      // please correct me, if I'm wrong, 

      // or comment next line for RFC2045 conformance, if you like

      if (!($bEmulate_imap_8bit && ($i==count($aLines)-1)))

          

      if (($iLastChar==0x09)||($iLastChar==0x20)) {

        $sLine{$iLength-1}='=';

        $sLine .= ($iLastChar==0x09)?'09':'20';

      }

    }    // imap_8bit encodes x20 before chr(13), too

    // although IMHO not requested by RFC2045, why not do it safer :)

    // and why not encode any x20 around chr(10) or chr(13)

    if ($bEmulate_imap_8bit) {

      $sLine=str_replace(' =0D','=20=0D',$sLine);

      //$sLine=str_replace(' =0A','=20=0A',$sLine);

      //$sLine=str_replace('=0D ','=0D=20',$sLine);

      //$sLine=str_replace('=0A ','=0A=20',$sLine);

    }



    // finally split into softlines no longer than 76 chars,

    // for even more safeness one could encode x09,x20 

    // at the very first character of the line 

    // and after soft linebreaks, as well,

    // but this wouldn't be caught by such an easy RegExp                   

    preg_match_all( '/.{1,73}([^=]{0,2})?/', $sLine, $aMatch );

    $sLine = implode( '=' . chr(13).chr(10), $aMatch[0] ); // add soft crlf's

  }



  // join lines into text

  return implode(chr(13).chr(10),$aLines);

}

?>

hans at lintoo dot dk
22-Dec-2005 06:05


The code below this note gives a lot of trouple with SpamAssasin, to fix it you will need to encode your subject with something like this:

<?php

/**

 * Function to encode a header if necessary

 * according to RFC2047

 */

private function encodeHeader($input, $charset = 'ISO-8859-1')

{

    preg_match_all('/(\s?\w*[\x80-\xFF]+\w*\s?)/', $input, $matches);

    foreach ($matches[1] as $value) {

        $replacement = preg_replace('/([\x20\x80-\xFF])/e', '"=" . strtoupper(dechex(ord("\1")))', $value);

        $input = str_replace($value, '=?' . $charset . '?Q?' . $replacement . '?=', $input);

    }

    

    return wordwrap($input,75,"\n\t",true);

}

?>

hans at lintoo dot dk
04-Oct-2005 10:35


I modified nick at plumdigitalmedia dot com's note so that it may support ISO-8859-1 encoded headers:

<?php

    public function encodeSubject($string,$prefix="=?ISO-8859-1?Q?",$postfix="?=")    {

        $crlf    = "\n\t";

        $string  = preg_replace('!(\r\n|\r|\n)!', $crlf, $string) . $crlf ;

        $f[]    = '/([\000-\010\013\014\016-\037\075\177-\377])/e' ;

        $r[]    = "'=' . sprintf('%02X', ord('\\1'))" ;

        $f[]    = '/([\011\040])' . $crlf . '/e' ;

        $r[]    = "'=' . sprintf('%02X', ord('\\1')) . '" . $crlf . "'" ;

        $string  = preg_replace($f, $r, $string);

        return $prefix.trim(wordwrap($string, 70 - strlen($prefix) - strlen($postfix), ' ' . $postfix . $crlf . $prefix, true)).$postfix;

    }

?>

Regards, Hans @ http://lintoo.dk/

marco dot manngatter at ecg-leipzig dot de
15-Sep-2005 03:43


you can use the following to make a larger subject.

use the imap_qprint() function, to convert your subject.



example:

<?php



$subject .= "=?iso-8859-1?Q?" . imap_qprint($subject) . "?=";



?>

nick at plumdigitalmedia dot com
10-May-2005 08:15


This function appears to wrap lines in the middle of words, not just at whitespace, which upsets some versions of Outlook Express when used to format email body text. We've had more luck with this function:



<?



    function quoted_printable($string)

    {

        $crlf    = "\n" ;

        $string  = preg_replace('!(\r\n|\r|\n)!', $crlf, $string) . $crlf ;

        $f[]     = '/([\000-\010\013\014\016-\037\075\177-\377])/e' ;

        $r[]     = "'=' . sprintf('%02X', ord('\\1'))" ;

        $f[]     = '/([\011\040])' . $crlf . '/e' ;

        $r[]     = "'=' . sprintf('%02X', ord('\\1')) . '" . $crlf . "'" ;

        $string  = preg_replace($f, $r, $string) ; 

        return trim(wordwrap($string, 70, ' =' . $crlf)) ;

    }



?>

k dot kozlowski at enter dot pl
14-Mar-2005 06:50


I had problems with encoding large subjects with polish characters. The problem was that imap_8bit() splits subject (on 75-th char) but when I add "=?ISO-8859-2?Q?" the header is too long.



This is a solution :

<?php

$subject = str_replace(" ", "_", trim($subject)) ;

// We need to delete "=\r\n" produced by imap_8bit() and replace '?'

$subject = str_replace("?", "=3F", str_replace("=\r\n", "", imap_8bit($subject))) ;

// Now we split by \r\n but with encoding text "=?ISO ..."

$subject = str_replace("\r\n", "?=\r\n =?ISO-8859-2?Q?", chunk_split($subject, 55)) ;

// We also have to remove last unneeded encoding text :

$subject = "=?ISO-8859-2?Q?" . substr($subject, 0, strlen($subject)-20) . "?=" ;

?>

rayFX
24-Jun-2004 09:14


imap_8bit seems to have a bug as it doesn't encode "?".

Had a lot of trouble to encode a mail-subject with

german umlaute with an ending "?"...



Try this:



$subject = "=?iso-8859-1?Q?" . str_replace("?","=3F",imap_8bit($tsubject_to_encode)) . "?=";

MagicalTux at FF.ST
15-Dec-2003 06:50


Reading Bully's note :

you can use the following to make a larger subject:

<?

$encoded = str_replace("=\r\n","",imap_8bit($string_to_encode)); 

?>

It's wrong. The header MAY NOT go over 75 chars per line.



The right solution :

<?

$encoded = rtrim( str_replace( "\n","\n\t", imap_8bit ($string_to_encode)))."\r\n";

?>



It will add \t at the start of the encoded new lines.

Mailers to see the lines after Subject: as extensions of this header.

php dot net at werner-ott dot de
09-Dec-2002 02:45


A comment on the "split-after-75-characters" phenomenon:



By splitting lines up after 75 characters, the function's behaviour is complying to RFC2047 (http://www.ietf.org/rfc/rfc2047.txt) (which specifies a protocol for the representation of non-ASCII text in message headers), section "2. Syntax of encoded-words".



A so called 'encoded word' has the following format:



   encoded-word = "=?" charset "?" encoding "?" encoded-text "?="



An 'encoded-word' may not be more than 75 characters long, including 'charset', 'encoding', 'encoded-text', and delimiters.  If it is desirable to encode more text than will fit in an 'encoded-word' of 75 characters, multiple 'encoded-word's (separated by CRLF SPACE) may be used.



hth

WOT

bully at newearth dot de
24-Apr-2002 07:48


you can use the following to make a larger subject:



$encoded = str_replace("=\r\n","",imap_8bit($string_to_encode));

zuffa at isdd dot sk
23-Sep-2000 08:57


Warning !


This function splits input text into


several lines aligned to 75 characters.


This is critical when you need input


text to be only one striaght line as


e.g. in e-mail header values.

imap_8bit

Description