When the client send Get data, utf-8 character encoding have a tiny problem with the urlencode.
Consider the "º" character.
Some clients can send (as example)
foo.php?myvar=%BA
and another clients send
foo.php?myvar=%C2%BA (The "right" url encoding)
in this scenary, you assign the value into variable $x
<?php
$x = $_GET['myvar'];
?>
$x store: in the first case "�" (bad) and in the second case "º" (good)
To fix that, you can use this function:
<?php
function to_utf8( $string ) {
// From http://w3.org/International/questions/qa-forms-utf-8.html
if ( preg_match('%^(?:
[\x09\x0A\x0D\x20-\x7E] # ASCII
| [\xC2-\xDF][\x80-\xBF] # non-overlong 2-byte
| \xE0[\xA0-\xBF][\x80-\xBF] # excluding overlongs
| [\xE1-\xEC\xEE\xEF][\x80-\xBF]{2} # straight 3-byte
| \xED[\x80-\x9F][\x80-\xBF] # excluding surrogates
| \xF0[\x90-\xBF][\x80-\xBF]{2} # planes 1-3
| [\xF1-\xF3][\x80-\xBF]{3} # planes 4-15
| \xF4[\x80-\x8F][\x80-\xBF]{2} # plane 16
)*$%xs', $string) ) {
return $string;
} else {
return iconv( 'CP1252', 'UTF-8', $string);
}
}
?>
and assign in this way:
<?php
$x = to_utf8( $_GET['myvar'] );
?>
$x store: in the first case "º" (good) and in the second case "º" (good)
Solve a lot of i18n problems.
Please fix the auto-urldecode of $_GET var in the next PHP version.
Bye.
Alejandro Salamanca
urldecode
(PHP 4, PHP 5, PHP 7, PHP 8)
urldecode — Decodes URL-encoded string
Description
Decodes any %##
encoding in the given string. Plus symbols ('+') are
decoded to a space character.
Parameters
-
string -
The string to be decoded.
Return Values
Returns the decoded string.
Examples
Example #1 urldecode() example
<?php
$query = "my=apples&are=green+and+red";
foreach (explode('&', $query) as $chunk) {
$param = explode("=", $chunk);
if ($param) {
printf("Value for parameter \"%s\" is \"%s\"<br/>\n", urldecode($param[0]), urldecode($param[1]));
}
}
?>Notes
See Also
- urlencode() - URL-encodes string
- rawurlencode() - URL-encode according to RFC 3986
- rawurldecode() - Decode URL-encoded strings
- » RFC 3986
add a note
User Contributed Notes 20 notes
alejandro at devenet dot net ¶
13 years ago
Visual ¶
18 years ago
If you are escaping strings in javascript and want to decode them in PHP with urldecode (or want PHP to decode them automatically when you're putting them in the query string or post request), you should use the javascript function encodeURIComponent() instead of escape(). Then you won't need any of the fancy custom utf_urldecode functions from the previous comments.
OZLperez11 at gmail dot com ¶
9 years ago
When sending a string via AJAX POST data which contains an ampersand (&), be sure to use encodeURIComponent() on the javascript side and use urldecode() on the php side for whatever variable that was. I've found it tricky to transfer raw ampersands and so this is what worked for me:
<?php
$_POST["data"] = "one%20%26%20two";
$a = urldecode($_POST["data"); // -> "one & two"
?>
For some reason, a variable with an ampersand would stay encoded while other POST variables were automatically decoded. I concatenated data from an html form before submitting, in case you wish to know what happened on the browser end.
tomas at penajaca dot com dot br ¶
21 years ago
urldecode does not decode "%0" bypassing it. I can cause troble when you are working with fixed lenght strings.
You can you the function below.
function my_urldecode($string){
$array = split ("%",$string);
if (is_array($array)){
while (list ($k,$v) = each ($array)){
$ascii = base_convert ($v,16,10);
$ret .= chr ($ascii);
}
}
return ("$ret");
}
Matt Johnson ¶
19 years ago
A reminder: if you are considering using urldecode() on a $_GET variable, DON'T!
Evil PHP:
<?php
# BAD CODE! DO NOT USE!
$term = urldecode($_GET['sterm']);
?>
Good PHP:
<?php
$term = $_GET['sterm'];
?>
The webserver will arrange for $_GET to have been urldecoded once already by the time it reaches you!
Using urldecode() on $_GET can lead to extreme badness, PARTICULARLY when you are assuming "magic quotes" on GET is protecting you against quoting.
Hint: script.php?sterm=%2527 [...]
PHP "receives" this as %27, which your urldecode() will convert to "'" (the singlequote). This may be CATASTROPHIC when injecting into SQL or some PHP functions relying on escaped quotes -- magic quotes rightly cannot detect this and will not protect you!
This "common error" is one of the underlying causes of the Santy.A worm which affects phpBB < 2.0.11.
rosty dot kerei at gmail dot com ¶
18 years ago
This function doesn't decode unicode characters. I wrote a function that does.
function unicode_urldecode($url)
{
preg_match_all('/%u([[:alnum:]]{4})/', $url, $a);
foreach ($a[1] as $uniord)
{
$dec = hexdec($uniord);
$utf = '';
if ($dec < 128)
{
$utf = chr($dec);
}
else if ($dec < 2048)
{
$utf = chr(192 + (($dec - ($dec % 64)) / 64));
$utf .= chr(128 + ($dec % 64));
}
else
{
$utf = chr(224 + (($dec - ($dec % 4096)) / 4096));
$utf .= chr(128 + ((($dec % 4096) - ($dec % 64)) / 64));
$utf .= chr(128 + ($dec % 64));
}
$url = str_replace('%u'.$uniord, $utf, $url);
}
return urldecode($url);
}
Jan Vratny ¶
16 years ago
mkaganer at gmail dot com:
try using encodeURI() instead of encode() in javascript. That worked for me, while your solution did not on __some__ national characters (at least in IE6).
bloodjazman at gmail dot com ¶
4 years ago
"+" replaced by space according to HTML x-www-form-url-encoded media type
see http://www.faqs.org/rfcs/rfc1866.html
aravind dot a dot padmanabhan at gmail dot com ¶
11 years ago
It seems that the $_REQUEST global parameter is automatically decoded only if the content type is application/x-www-form-urlencoded.
if the content type is multipart/form-data. the data remains un-decoded. and we have to manually handle the decoding at our end
Joe ¶
16 years ago
It's worth pointing out that if you are using AJAX and need to encode strings that are being sent to a PHP application, you may not need to decode them in PHP.
<?php
echo stripslashes(nl2br($_POST['message']));
?>
Will properly output a message sent with the javascript code if the message is encoded:
message = encodeURIComponent(message)
And is sent with an AJAX POST request with the header:
ajaxVar.setRequestHeader('Content-type', 'application/x-www-form-urlencoded')
smolniy at mtu dot ru ¶
21 years ago
For compatibility of new and old brousers:
%xx -> char
%u0xxxx -> char
function unicode_decode($txt) {
$txt = ereg_replace('%u0([[:alnum:]]{3})', '&#x\1;',$txt);
$txt = ereg_replace('%([[:alnum:]]{2})', '&#x\1;',$txt);
return ($txt);
}
mkaganer at gmail dot com ¶
16 years ago
B.H.
I had troubles converting Unicode-encoded data in $_GET (like this: %u05D8%u05D1%u05E2) which is generated by JavaScript's escape() function to UTF8 for server-side processing.
Finally, i've found a simple solution (only 3 lines of code) that does it (at least in my configuration):
<?php
function utf8_urldecode($str) {
$str = preg_replace("/%u([0-9a-f]{3,4})/i","&#x\\1;",urldecode($str));
return html_entity_decode($str,null,'UTF-8');;
}
?>
note that documentation for html_entity_decode() states that "Support for multi-byte character sets was added at PHP 5.0.0" so this might not work for PHP 4
Soaku ¶
7 years ago
Even if $_GET and $_REQUEST are decoded automatically, some other variables aren't. For example $_SERVER["REQUEST_URI"]. Remember to decode it when using.
spam at soiland dot no ¶
19 years ago
About reg_var and "html reserved words"
Do not add spaces as the user suggests.
Instead, do what all HTML standards says and encode & in URLs as & in your HTML.
The reason why & works "most of the time" is that browsers are forgiving and just decode the & as the &-sign. This breaks whenever you have a variable that matches an HTML entity, like "gt" or "copy" or whatever. © in your URL will be interpreted as © (the ; is not mandatory in SGML as it is "implied". In XML it is mandatory.). The result will be the same as if you had inserted the actual character into your source code, for instance by pressing alt-0169 and actually inserted ? in your HTML.
Ie, use:
<a href="?name=stain&fish=knott">mylink</a>
Note that the decoding of & to & is done in the browser, and it's done right after splitting the HTML into tags, attributes and content, but it works both for attributes and content.
This mean you should &entitify all &-s in any other HTML attributes as well, such as in a form with
<input name="fish" value="fish & fries" />.
regindk at hotmail dot com ¶
21 years ago
About: bellani at upgrade4 dot it
$str = "pippo.php?param1=®_var";
echo rawurldecode($str);
Gives:
pippo.php?param1=?_var
Instead of using a space you should exchange & with the correct W3C &
Like this:
$str = "pippo.php?param1=&reg_var";
echo rawurldecode($str);
mail dot roliveira at gmail dot com ¶
15 years ago
Send json to PHP via AJAX (POST)
If you send json data via ajax, and encode it with encodeURIComponent in javascript, then on PHP side, you will have to do stripslashes on your $_POST['myVar'].
After this, you can do json_decode on your string.
Ex.:
<?php
// first use encodeURIComponent on javascript to encode the string
// receive json string and prepare it to json_decode
$jsonStr = stripslashes ($_POST['action']);
// decode to php object
$json = json_decode ($jsonStr);
// $json is now a php object
?>
bellani at upgrade4 dot it ¶
21 years ago
If you have a "html reserved word" as variable name (i.e. "reg_var") and you pass it as an argument you will get a wrong url. i.e.
<a href="pippo.php?param1=®_var=">go</a>
you will get a wrong url like this
"pippo.php?param1=?_var"
Simply add a space between "&" and "reg_var" and it will work!
<a href="pippo.php?param1=& reg_var=">go</a>
"pippo.php?param1=&%20reg_var"
Works!!
igjav at cesga dot es ¶
22 years ago
This seems to decode correctly between most browsers and charater coding configurations. Specially indicated for direct parsing of URL as it comes on environment variables:
function crossUrlDecode($source) {
$decodedStr = '';
$pos = 0;
$len = strlen($source);
while ($pos < $len) {
$charAt = substr ($source, $pos, 1);
if ($charAt == '?') {
$char2 = substr($source, $pos, 2);
$decodedStr .= htmlentities(utf8_decode($char2),ENT_QUOTES,'ISO-8859-1');
$pos += 2;
}
elseif(ord($charAt) > 127) {
$decodedStr .= "&#".ord($charAt).";";
$pos++;
}
elseif($charAt == '%') {
$pos++;
$hex2 = substr($source, $pos, 2);
$dechex = chr(hexdec($hex2));
if($dechex == '?') {
$pos += 2;
if(substr($source, $pos, 1) == '%') {
$pos++;
$char2a = chr(hexdec(substr($source, $pos, 2)));
$decodedStr .= htmlentities(utf8_decode($dechex . $char2a),ENT_QUOTES,'ISO-8859-1');
}
else {
$decodedStr .= htmlentities(utf8_decode($dechex));
}
}
else {
$decodedStr .= $dechex;
}
$pos += 2;
}
else {
$decodedStr .= $charAt;
$pos++;
}
}
return $decodedStr;
}
caribe at flash-brasil dot com dot br ¶
21 years ago
To allow urldecode to work with Brazilian characters as ? ? ? and other just place this header command :
header('Content-type: text/html; charset=UTF-8');
Anonymous ¶
21 years ago
nataniel, your function needs to be corrected as follows:
------------------------------------------------------------
function unicode_decode($txt) {
return ereg_replace('%u([[:alnum:]]{4})', '&#x\1;',$txt);
}
------------------------------------------------------------
since some codes does not begin with %u0.
