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**.
quoted_printable_decode
(PHP 4, PHP 5, PHP 7, PHP 8)
quoted_printable_decode — Convierte un string quoted-printable en un string de 8 bits
Descripción
Esta función devuelve un string binario de 8 bits que corresponde al string quoted printable decodificado (de acuerdo con el » RFC2045, sección 6.7, no el » RFC2821, sección 4.5.2, así que puntos adicionales no son retirados del principio de línea).
Esta función es similar a imap_qprint(), excepto que no requiere del módulo IMAP para funcionar.
Parámetros
-
str -
El string de entrada.
Valores devueltos
Devuelve el string binario de 8 bits.
add a note
User Contributed Notes 21 notes
jab_creations at yahoo dot com ¶
3 years ago
madmax at express dot ru ¶
24 years ago
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
Karora ¶
16 years ago
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.
legolas558 ¶
17 years ago
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
Thomas Pequet / Memotoo.com ¶
18 years ago
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 ?"
soletan at toxa dot de ¶
17 years ago
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.
pob at medienrecht dot NOSPAM dot org ¶
23 years ago
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;
zg ¶
16 years ago
<?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;
}
?>
ludwig at gramberg-webdesign dot de ¶
17 years ago
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 ¶
17 years ago
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")
: $return;
}
?>
sven at e7o dot de ¶
4 years ago
If you're really lazy and producing HTML anyways and the end, just convert it to HTML entities and move the Unicode/ISO struggling to the document's encoding:
<?php
function qpd($e)
{
return preg_replace_callback(
'/=([a-z0-9]{2})/i',
function ($m) {
return '�' . $m[1] . ';';
},
$e
);
}
?>
yual at inbox dot ru ¶
11 years ago
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);}
feedr ¶
15 years ago
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 $string;
}
}
?>
andre at luyer dot nl ¶
16 years ago
A small update for Andrew's code below. This one leaves the original CRLF pairs intact (and allowing the preg_replace to work as intended):
<?php
if (!function_exists("quoted_printable_encode")) {
/**
* Process a string to fit the requirements of RFC2045 section 6.7. Note that
* this works, but replaces more characters than the minimum set. For readability
* the spaces and CRLF pairs aren't encoded though.
*/
function quoted_printable_encode($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
Christian Albrecht ¶
16 years ago
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 $tmp;
}
?>
david at lionhead dot nl ¶
16 years ago
My version of quoted_printable encode, as the convert.quoted-printable-encode filter breaks on outlook express. This one seems to work on express/outlook/thunderbird/gmail.
function quoted_printable_encode($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;
}
vita dot plachy at seznam dot cz ¶
16 years ago
<?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*/);
legolas558 at users dot sausafe dot net ¶
17 years ago
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 ;)
?>
steffen dot weber at computerbase dot de ¶
19 years ago
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().
naitsirch at e dot mail dot de ¶
4 years ago
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.");
// $result === 'This is a'
?>
This is not a bug, but the intended behaviour and defined by RFC 2045 (see https://www.ietf.org/rfc/rfc2045.txt) in paragraph 2.7 and 2.8.
dmitry at koterov dot ru ¶
19 years ago
Previous comment has a bug: encoding of short test does not work because of incorrect usage of preg_match_all(). Have somebody read it at all? :-)
Correct version (seems), with additional imap_8bit() function emulation:
if (!function_exists('imap_8bit')) {
function imap_8bit($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;
}
