(PHP 4, PHP 5)
quoted_printable_decode — Convert a quoted-printable string to an 8 bit string
$str
)This function returns an 8-bit binary string corresponding to the decoded quoted printable string (according to » RFC2045, section 6.7, not » RFC2821, section 4.5.2, so additional periods are not stripped from the beginning of line).
This function is similar to imap_qprint(), except this one does not require the IMAP module to work.
str
The input string.
Returns the 8-bit binary string.
feedr (2009-03-06 14:39:37)
Another (improved) version of quoted_printable_encode(). Please note the order of the array elements in str_replace().
I've just rewritten the previous function for better readability.
<?php
if (!function_exists("quoted_printable_encode")) {
/**
* Process a string to fit the requirements of RFC2045 section 6.7. Note that
* this works, but replaces more characters than the minimum set. For readability
* the spaces and CRLF pairs aren't encoded though.
*/
function quoted_printable_encode($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 $string;
}
}
?>
andre at luyer dot nl (2008-11-02 12:36:51)
A small update for Andrew's code below. This one leaves the original CRLF pairs intact (and allowing the preg_replace to work as intended):
<?php
if (!function_exists("quoted_printable_encode")) {
/**
* Process a string to fit the requirements of RFC2045 section 6.7. Note that
* this works, but replaces more characters than the minimum set. For readability
* the spaces and CRLF pairs aren't encoded though.
*/
function quoted_printable_encode($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
Karora (2008-10-19 20:27:15)
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.
zg (2008-09-27 06:48:09)
<?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 $enc;
}
?>
Christian Albrecht (2008-09-24 15:10:25)
In Addition to david lionhead's function:
<?php
function quoted_printable_encode($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 $tmp;
}
?>
lewiswright [ at t] gmail [d dot t] com (2008-09-24 11:16:09)
In response to Bobo, using:
chunk_split($string, 76, "=\n");
Can result to broken encoding (in some rare cases). This is for a couple of reasons. Firstly, the chunk split could insert a newline between an encoding if it was on the end of a line. So it could turn an entity: =20 in to =2=\n0 - which breaks it. Secondly, the spec says that the line should be no longer than 76 in total, which includes the '=' used to indicate a soft newline.
Additionally, there's a couple of minor readability issues with it. Firstly, chunk_split leaves a trailing newline - which is unnecessary. Lastly, it does not consider newlines already in the text (which breaks readability) so it often wraps when not needed.
Here's a snippet which will fix all the issues mentioned:
preg_replace('/[^\r\n]{73}[^=\r\n]{2}/', "$0=\r\n", $data);
david at lionhead dot nl (2008-08-18 05:54:48)
My version of quoted_printable encode, as the convert.quoted-printable-encode filter breaks on outlook express. This one seems to work on express/outlook/thunderbird/gmail.
function quoted_printable_encode($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 $tmp;
}
Bobo (2008-07-12 09:20:22)
Following up to Hoffa's function, return the result like:
chunk_split($string, 76, "=\n");
to conform to the RFC, which requires that lines are no longer than 76 chars.
h atnospam hoffa dotelidot se (2008-03-31 17:32:10)
I do like this to encode
function quoted_printable_encode($string) {
$string = rawurlencode($string);
$string = str_replace("%","=",$string);
RETURN $string;
}
best regards / Hoffa
vita dot plachy at seznam dot cz (2008-03-23 08:22:46)
<?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*/);
ludwig at gramberg-webdesign dot de (2007-09-24 12:13:03)
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;
}
?>
roelof (2007-07-24 07:06:29)
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;
}
?>
legolas558 at users dot sausafe dot net (2007-02-21 11:09:07)
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?'.$result.'?=';
}
I hope it helps ;)
?>
soletan at toxa dot de (2007-02-20 15:54:02)
Be warned! The method below for encoding text does not work as requested by RFC1521!
Consider a line consisting of 75 'A' and a single é (or similar non-ASCII character) ... the method below would encode and return a line of 78 octets, breaking with RFC 1521, 5.1 Rule #5: "The Quoted-Printable encoding REQUIRES that encoded lines be no more than 76 characters long."
Good QP-encoding takes a bit more than this.
legolas558 (2007-02-12 10:32:59)
Please note that in the below encode function there is a bug!
<?php
if (($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
MagicalTux at ookoo dot org (2007-02-06 03:44:15)
I checked Thomas Pequet / Memotoo.com's quoted_printable_encode() function, and it's rather non-respective to standards...
Here's my own, fixed version :
<?php
function quoted_printable_encode($string, $linelen = 0, $linebreak="=\r\n", $breaklen = 0, $encodecrlf = false) {
// Quoted printable encoding is rather simple.
// Each character in the string $string should be encoded if:
// Character code is <0x20 (space)
// Character is = (as it has a special meaning: 0x3d)
// Character is over ASCII range (>=0x80)
$len = strlen($string);
$result = '';
for($i=0;$i<$len;$i++) {
if ($linelen >= 76) { // break lines over 76 characters, and put special QP linebreak
$linelen = $breaklen;
$result.= $linebreak;
}
$c = ord($string[$i]);
if (($c==0x3d) || ($c>=0x80) || ($c<0x20)) { // in this case, we encode...
if ((($c==0x0A) || ($c==0x0D)) && (!$encodecrlf)) { // but not for linebreaks
$result.=chr($c);
$linelen = 0;
continue;
}
$result.='='.str_pad(strtoupper(dechex($c)), 2, '0');
$linelen += 3;
continue;
}
$result.=chr($c); // normal characters aren't encoded
$linelen++;
}
return $result;
}
$test = 'This is a test !! héhéhé ....'."\r\n";
$test.= 'You can write really long pieces of text without having to worry about mail transfert. Quoted printable supports special way to handle such lines, by instering an escaped linebreak. This linebreak, once converted back to 8bit, will disappear.';
echo quoted_printable_encode(utf8_encode($test))."\r\n";
?>
By default, the function encodes text to be used as a quoted-printable-encoded body, however by tweaking the parameters, you can get subject-header encoding and stuff like that.
Thomas Pequet / Memotoo.com (2006-10-19 06:15:38)
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 ?"
umu (2006-02-06 14:38:40)
...deed.ztinmehc-ut.zrh@umuumu@hrz.tu-chemnitz.deed...
and write [^=]{0,2} in the last regular expression
to avoid the soft linebreak "=" being the 77th char in line,
see my imap_8bit() emulation at
http://www.php.net/manual/en/function.imap-8bit.php#61216
steffen dot weber at computerbase dot de (2005-07-22 14:08:09)
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().
bendi at interia dot pl (2005-03-29 01:08:22)
This function appeared twice already and in both versions it's got error in rule replacing chunk_split, because it misses the last line in encoded text if it's shorter than 73 characters.
My solution (got rid of replace_callback)
<?
function quoted_printable_encode( $sString ) {
/*instead of replace_callback i used <b>e</b> modifier for regex rule, which works as eval php function*/
$sString = preg_replace( '/[^\x21-\x3C\x3E-\x7E\x09\x20]/e', 'sprintf( "=%02x", ord ( "$0" ) ) ;', $sString );
/*now added to this rule one or more chars which lets last line to be matched and included in results*/
preg_match_all( '/.{1,73}([^=]{0,3})?/', $sString, $aMatch );
return implode( '=' . CR, $aMatch[0] );
}
?>
dmitry at koterov dot ru (2005-02-19 05:26:50)
Previous comment has a bug: encoding of short test does not work because of incorrect usage of preg_match_all(). Have somebody read it at all? :-)
Correct version (seems), with additional imap_8bit() function emulation:
if (!function_exists('imap_8bit')) {
function imap_8bit($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;
}
drm at melp dot nl (2005-02-09 03:32:34)
A easier, improved way of encoding for quoted-printable transfer:
------
function quoted_printable_encode_character ( $matches ) {
$character = end ( $matches );
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)
preg_match_all ( '/.{73}([^=]{0,3})/', $string, $match ); // Rule #1
return implode ( $newline, $match[0] );
}
-----
tamas dot tompa at kirowski dot com (2004-02-20 07:46:15)
i have found a bug in pob at medienrecht dot NOSPAM dot org
's qp_enc function. in quoted printable messages need to convert the first pont in the lines too...
here is the fixed code:
function qp_enc( $input = "", $line_max = 76, $space_conv = false ) {
$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 ( ( $i == 0 ) && ( $dec == 46 ) ) { //
convert first point in the line into =2E
$c = "=2E";
}
if ( $dec == 32 ) {
if ( $i == ( $linlen - 1 ) ) { // convert
space at eol only
$c = "=20";
} else if ( $space_conv ) {
$c = "=20";
}
} elseif ( ($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 = "";
// check if newline first character will
be point or not
if ( $dec == 46 ) {
$c = "=2E";
}
}
$newline .= $c;
} // end of for
$output .= $newline.$eol;
} // end of while
return trim($output);
}
pob at medienrecht dot NOSPAM dot org (2001-07-18 13:06:09)
madmax at express dot ru (2000-08-03 17:41:08)
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
gustavf at spamstop dot com (2000-07-07 10:25:31)
This function does not recognize lowercase letters as part of the Quoted Printable encoding. RFC 1521 specifies:
Uppercase letters must be used when sending hexadecimal data, though a robust implementation may choose to recognize lowercase letters on receipt.