If you're getting black diamonds or weird characters that seemingly block an echo but still encounter strlen($string) > 0 you're probably encountering an encoding issue. Unlike the people writing ENCODE functions on a DECODE page I will actually talk about DECODE on a DECODE page.

The specific problem I encountered was that an email was encoded using a Russian encoding (KOI8-R) though I output everything as UTF-8 because: compatibility.

If you try to do this with a Russian encoding:

<?php
echo quoted_printable_decode('=81');
?>

You'll get that corrupted data.

I did a couple of tests and it turns out the following is how you nest the mb_convert_encoding function:

<?php
echo '<p>Test: "'.mb_convert_encoding(quoted_printable_decode('=81'), 'UTF-8', 'KOI8-R').'".</p>';
?>

Unfortunately I could not find a character mapping table or anything listed under RFC 2045 Section 6.7. However I came across the website https://dencode.com/en/string/quoted-printable which allows you to manually choose the encoding (it's an open source site, they have a GIT repository for the morbidly curious).

As it turns out the start of the range is relative to the encoding. So Latin (ISO-8859-1) and Russian (KOI8-R) will likely (not tested this) encode to different characters **relative to the string encoding**.

Taking a bunch of the earlier comments together, you can synthesize a nice short and reasonably efficient quoted_printable_encode function like this:

Note that I put this in my standard library file, so I wrap it in a !function_exists in order that if there is a pre-existing PHP one it will just work and this will evaluate to a noop.

<?php
if ( !function_exists("quoted_printable_encode") ) {
  /**
  * Process a string to fit the requirements of RFC2045 section 6.7.  Note that
  * this works, but replaces more characters than the minimum set. For readability
  * the spaces aren't encoded as =20 though.
  */
  function quoted_printable_encode($string) {
    return preg_replace('/[^\r\n]{73}[^=\r\n]{2}/', "$0=\r\n", str_replace("%","=",str_replace("%20"," ",rawurlencode($string))));
  }
}
?>

Regards,
Andrew McMillan.

Be warned! The method below for encoding text does not work as requested by RFC1521!

Consider a line consisting of 75 'A' and a single é (or similar non-ASCII character) ... the method below would encode and return a line of 78 octets, breaking with RFC 1521, 5.1 Rule #5: "The Quoted-Printable encoding REQUIRES that encoded lines be no more than 76 characters long."

Good QP-encoding takes a bit more than this.

<?php

function quoted_printable_encode( $str, $chunkLen = 72 )
{
    $offset = 0;
    
    $str = strtr(rawurlencode($str), array('%' => '='));
    $len = strlen($str);
    $enc = '';
    
    while ( $offset < $len )
    {
        if ( $str{ $offset + $chunkLen - 1 } === '=' )
        {
            $line = substr($str, $offset, $chunkLen - 1);
            $offset += $chunkLen - 1;
        }
        elseif ( $str{ $offset + $chunkLen - 2 } === '=' )
        {
            $line = substr($str, $offset, $chunkLen - 2);
            $offset += $chunkLen - 2;
        }
        else 
        {
            $line = substr($str, $offset, $chunkLen);
            $offset += $chunkLen;
        }
        
        if ( $offset + $chunkLen < $len )
            $enc .= $line ."=\n";
        else 
            $enc .= $line;
    }
    
    return $enc;
}

?>

Some  browser (netscape, for example)
send 8-bit quoted printable text like this:
=C5=DD=A3=D2=C1= =DA

"= =" means continuos word.
 php function not detect this situations and translate in string like:
 abcde=f

If you want a function to do the reverse of "quoted_printable_decode()", follow the link you will find the "quoted_printable_encode()" function:
http://www.memotoo.com/softs/public/PHP/quoted printable_encode.inc.php

Compatible "ENCODING=QUOTED-PRINTABLE"
Example: 
quoted_printable_encode(ut8_encode("c'est quand l'été ?")) 
-> "c'est quand l'=C3=A9t=C3=A9 ?"

If you're really lazy and producing HTML anyways and the end, just convert it to HTML entities and move the Unicode/ISO struggling to the document's encoding:

<?php
function qpd($e)
{
    return preg_replace_callback(
        '/=([a-z0-9]{2})/i',
        function ($m) {
            return '&#x00' . $m[1] . ';';
        },
        $e
    );
}
?>

I use a hack for this bug:

$str = str_replace("=\r\n", '', quoted_printable_encode($str));
if (strlen($str) > 73) {$str = substr($str,0,74)."=\n".substr($str,74);}

As the two digit hexadecimal representation SHOULD be in uppercase you want to use "=%02X" (uppercase X) instead of "=%02x" as the first argument to sprintf().

If you do not have access to imap_* and do not want to use 
?$message = chunk_split( base64_encode($message) );?
because you want to be able to read the ?source? of your mails, you might want to try this:
(any suggestions very welcome!)


function qp_enc($input = "quoted-printable encoding test string", $line_max = 76) {

    $hex = array('0','1','2','3','4','5','6','7','8','9','A','B','C','D','E','F');
    $lines = preg_split("/(?:\r\n|\r|\n)/", $input);
    $eol = "\r\n";
    $escape = "=";
    $output = "";

    while( list(, $line) = each($lines) ) {
        //$line = rtrim($line); // remove trailing white space -> no =20\r\n necessary
        $linlen = strlen($line);
        $newline = "";
        for($i = 0; $i < $linlen; $i++) {
            $c = substr($line, $i, 1);
            $dec = ord($c);
            if ( ($dec == 32) && ($i == ($linlen - 1)) ) { // convert space at eol only
                $c = "=20"; 
            } elseif ( ($dec == 61) || ($dec < 32 ) || ($dec > 126) ) { // always encode "\t", which is *not* required
                $h2 = floor($dec/16); $h1 = floor($dec%16); 
                $c = $escape.$hex["$h2"].$hex["$h1"]; 
            }
            if ( (strlen($newline) + strlen($c)) >= $line_max ) { // CRLF is not counted
                $output .= $newline.$escape.$eol; // soft line break; " =\r\n" is okay
                $newline = "";
            }
            $newline .= $c;
        } // end of for
        $output .= $newline.$eol;
    }
    return trim($output);

}

$eight_bit = "\xA7 \xC4 \xD6 \xDC \xE4 \xF6 \xFC \xDF         =          xxx             yyy       zzz          \r\n"
            ." \xA7      \r \xC4 \n \xD6 \x09    "; 
print $eight_bit."\r\n---------------\r\n";
$encoded = qp_enc($eight_bit); 
print $encoded;

Please note that in the below encode function there is a bug!

<?php
if (($c==0x3d) || ($c>=0x80) || ($c<0x20))
?>

$c should be checked against less or equal to encode spaces!

so the correct code is

<?php
if (($c==0x3d) || ($c>=0x80) || ($c<=0x20))
?>

Fix the code or post this note, please

A small update for Andrew's code below. This one leaves the original CRLF pairs intact (and allowing the preg_replace to work as intended):

<?php
if (!function_exists("quoted_printable_encode")) {
  /**
  * Process a string to fit the requirements of RFC2045 section 6.7. Note that
  * this works, but replaces more characters than the minimum set. For readability
  * the spaces and CRLF pairs aren't encoded though.
  */
  function quoted_printable_encode($string) {
    return preg_replace('/[^\r\n]{73}[^=\r\n]{2}/', "$0=\r\n",
      str_replace("%", "=", str_replace("%0D%0A", "\r\n", 
        str_replace("%20"," ",rawurlencode($string)))));
  }
}
?>

Regards, André

Another (improved) version of quoted_printable_encode(). Please note the order of the array elements in str_replace().
I've just rewritten the previous function for better readability.

<?php
if (!function_exists("quoted_printable_encode")) {
  /**
  * Process a string to fit the requirements of RFC2045 section 6.7. Note that
  * this works, but replaces more characters than the minimum set. For readability
  * the spaces and CRLF pairs aren't encoded though.
  */
  function quoted_printable_encode($string) {
        $string = str_replace(array('%20', '%0D%0A', '%'), array(' ', "\r\n", '='), rawurlencode($string));
        $string = preg_replace('/[^\r\n]{73}[^=\r\n]{2}/', "$0=\r\n", $string);

        return $string;
  }
}
?>

In Addition to david lionhead's function:

<?php
function quoted_printable_encode($txt) {
    /* Make sure there are no %20 or similar */
    $txt = rawurldecode($txt);
    $tmp="";
    $line="";
    for ($i=0;$i<strlen($txt);$i++) {
        if (($txt[$i]>='a' && $txt[$i]<='z') || ($txt[$i]>='A' && $txt[$i]<='Z') || ($txt[$i]>='0' && $txt[$i]<='9')) {
            $line.=$txt[$i];
            if (strlen($line)>=75) {
                $tmp.="$line=\n";
                $line="";
            }
        }
        else {
            /* Important to differentiate this case from the above */
            if (strlen($line)>=72) {
            $tmp.="$line=\n";
            $line="";
            }
            $line.="=".sprintf("%02X",ord($txt[$i]));
        }
    }
    $tmp.="$line\n";
    return $tmp;
}
?>

My version of quoted_printable encode, as the convert.quoted-printable-encode filter breaks on outlook express. This one seems to work on express/outlook/thunderbird/gmail.

function quoted_printable_encode($txt) {
    $tmp="";
    $line="";
    for ($i=0;$i<strlen($txt);$i++) {
        if (($txt[$i]>='a' && $txt[$i]<='z') || ($txt[$i]>='A' && $txt[$i]<='Z') || ($txt[$i]>='0' && $txt[$i]<='9'))
            $line.=$txt[$i];
        else
            $line.="=".sprintf("%02X",ord($txt[$i]));
        if (strlen($line)>=75) {
            $tmp.="$line=\n";
            $line="";
        }
    }
    $tmp.="$line\n";
    return $tmp;
}

<?php
$text = <<<EOF
This function enables you to convert text to a quoted-printable string as well as to create encoded-words used in email headers (see http://www.faqs.org/rfcs/rfc2047.html).

No line of returned text will be longer than specified. Encoded-words will not contain a newline character. Special characters are removed.
EOF;

define('QP_LINE_LENGTH', 75);
define('QP_LINE_SEPARATOR', "\r\n");

function quoted_printable_encode($string, $encodedWord = false)
{
    if(!preg_match('//u', $string)) {
        throw new Exception('Input string is not valid UTF-8');
    }
    
    static $wordStart = '=?UTF-8?Q?';
    static $wordEnd = '?=';
    static $endl = QP_LINE_SEPARATOR;
    
    $lineLength = $encodedWord
        ? QP_LINE_LENGTH - strlen($wordStart) - strlen($wordEnd)
        : QP_LINE_LENGTH;
    
    $string = $encodedWord
        ? preg_replace('~[\r\n]+~', ' ', $string)    // we need encoded word to be single line
        : preg_replace('~\r\n?~', "\n", $string);    // normalize line endings
    $string = preg_replace('~[\x00-\x08\x0B-\x1F]+~', '', $string);    // remove control characters
    
    $output = $encodedWord ? $wordStart : '';
    $charsLeft = $lineLength;
    
    $chr = isset($string{0}) ? $string{0} : null;
    $ord = ord($chr);
    
    for ($i = 0; isset($chr); $i++) {
        $nextChr = isset($string{$i + 1}) ? $string{$i + 1} : null;
        $nextOrd = ord($nextChr);
        
        if (
            $ord > 127 or    // high byte value
            $ord === 95 or    // underscore "_"
            $ord === 63 && $encodedWord or    // "?" in encoded word
            $ord === 61 or    // equal sign "="
            // space or tab in encoded word or at line end
            $ord === 32 || $ord === 9 and $encodedWord || !isset($nextOrd) || $nextOrd === 10
        ) {
            $chr = sprintf('=%02X', $ord);    
        }
        
        if ($ord === 10) {    // line feed
            $output .= $endl;
            $charsLeft = $lineLength;
        } elseif (
            strlen($chr) < $charsLeft or
            strlen($chr) === $charsLeft and $nextOrd === 10 || $encodedWord
        ) {    // add character
            $output .= $chr;
            $charsLeft-=strlen($chr);
        } elseif (isset($nextOrd)) {    // another line needed
            $output .= $encodedWord
                ? $wordEnd . $endl . "\t" . $wordStart . $chr
                : '=' . $endl . $chr;
            $charsLeft = $lineLength - strlen($chr);
        }
        
        $chr = $nextChr;
        $ord = $nextOrd;
    }
    
    return $output . ($encodedWord ? $wordEnd : '');
}

echo quoted_printable_encode($text/*, true*/);

my approach for quoted printable encode using the stream converting abilities

<?php
/**
 * @param string $str
 * @return string
 * */
function quoted_printable_encode($str) {
    $fp = fopen('php://temp', 'w+');
    stream_filter_append($fp, 'convert.quoted-printable-encode');
    fwrite($fp, $str);    
    fseek($fp, 0);
    $result = '';
    while(!feof($fp))
        $result .= fread($fp, 1024);
    fclose($fp);
    return $result;
}
?>

I modified the below version of legolas558 at users dot sausafe dot net and added a wrapping option.

<?php
/**
 *    Codeer een String naar zogenaamde 'quoted printable'. Dit type van coderen wordt
 *    gebruikt om de content van 8 bit e-mail berichten als 7 bits te versturen.
 *
 *    @access public
 *    @param string    $str    De String die we coderen
 *    @param bool      $wrap   Voeg linebreaks toe na 74 tekens?
 *    @return string
 */

function quoted_printable_encode($str, $wrap=true)
{
    $return = '';
    $iL = strlen($str);
    for($i=0; $i<$iL; $i++)
    {
        $char = $str[$i];
        if(ctype_print($char) && !ctype_punct($char)) $return .= $char;
        else $return .= sprintf('=%02X', ord($char));
    }
    return ($wrap === true)
        ? wordwrap($return, 74, " =\n")
        : $return;
}

?>

As soletan at toxa dot de reported, that function is very bad and does not provide valid enquoted printable strings. While using it I saw spam agents marking the emails as QP_EXCESS and sometimes the email client did not recognize the header at all; I really lost time :(. This is the new version (we use it in the Drake CMS core) that works seamlessly:

<?php

//L: note $encoding that is uppercase
//L: also your PHP installation must have ctype_alpha, otherwise write it yourself
function quoted_printable_encode($string, $encoding='UTF-8') {
// use this function with headers, not with the email body as it misses word wrapping
       $len = strlen($string);
       $result = '';
       $enc = false;
       for($i=0;$i<$len;++$i) {
        $c = $string[$i];
        if (ctype_alpha($c))
            $result.=$c;
        else if ($c==' ') {
            $result.='_';
            $enc = true;
        } else {
            $result.=sprintf("=%02X", ord($c));
            $enc = true;
        }
       }
       //L: so spam agents won't mark your email with QP_EXCESS
       if (!$enc) return $string;
       return '=?'.$encoding.'?q?'.$result.'?=';
}

I hope it helps ;)

?>

Previous comment has a bug: encoding of short test does not work because of incorrect usage of preg_match_all(). Have somebody read it at all? :-)

Correct version (seems), with additional imap_8bit() function emulation:

if (!function_exists('imap_8bit')) {
 function imap_8bit($text) {
   return quoted_printable_encode($text);
 }
}

function quoted_printable_encode_character ( $matches ) {
   $character = $matches[0];
   return sprintf ( '=%02x', ord ( $character ) );
}

// based on http://www.freesoft.org/CIE/RFC/1521/6.htm
function quoted_printable_encode ( $string ) {
   // rule #2, #3 (leaves space and tab characters in tact)
   $string = preg_replace_callback (
     '/[^\x21-\x3C\x3E-\x7E\x09\x20]/',
     'quoted_printable_encode_character',
     $string
   );
   $newline = "=\r\n"; // '=' + CRLF (rule #4)
   // make sure the splitting of lines does not interfere with escaped characters
   // (chunk_split fails here)
   $string = preg_replace ( '/(.{73}[^=]{0,3})/', '$1'.$newline, $string);
   return $string;
}

If there is a NULL byte in the string that is passed, quoted_printable_decode will crop everything after the NULL byte and the NULL byte itself.

<?php
$result = quoted_printable_decode("This is a\0 test.");
// $result === 'This is a'
?>

This is not a bug, but the intended behaviour and defined by RFC 2045 (see https://www.ietf.org/rfc/rfc2045.txt) in paragraph 2.7 and 2.8.