quoted_printable_decode()

(PHP 4, PHP 5, PHP 7)

将 quoted-printable 字符串转换为 8-bit 字符串

说明

quoted_printable_decode(string $str): string

该函数返回 quoted-printable 解码之后的 8-bit 字符串(参考» RFC2045 的6.7章节，而不是» RFC2821 的4.5.2章节，so additional periods are not stripped from the beginning of line)

该函数与 imap_qprint()函数十分相似，但是该函数不需要依赖 IMAP 模块。

参数

$str

输入的字符串。

返回值

返回的 8-bit 二进制字符串。

参见

• quoted_printable_encode()将 8-bit 字符串转换成 quoted-printable 字符串

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 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 ;)
?>

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

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.

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;
}
?>

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
$text = <<<EOF
This function enables you to convert text to a quoted-printable string as well as to create encoded-words used in email headers (see http://www.faqs.org/rfcs/rfc2047.html).
No line of returned text will be longer than specified. Encoded-words will not contain a newline character. Special characters are removed.
EOF;
define('QP_LINE_LENGTH', 75);
define('QP_LINE_SEPARATOR', "\r\n");
function quoted_printable_encode($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*/);

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 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;

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;
}

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

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
  );
}
?>

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;
 }
}
?>

<?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;
}
?>

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;
}

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;
}
?>

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().

I do like this to encode
function quoted_printable_encode($string) {
 $string = rawurlencode($string);
 $string = str_replace("%","=",$string);
 RETURN $string;
}
best regards / Hoffa

Following up to Hoffa's function, return the result like:
chunk_split($string, 76, "=\n");
to conform to the RFC, which requires that lines are no longer than 76 chars.