PHP IDE :: NuSphere Forums :: View topic - How to force UTF-8 encoding?

Posted: Tue Aug 15, 2006 3:00 am

Hi,

when requesting the wsdl description page from my nusoap server the returned xml is ISO-8859-1 encoded. How can i force it to be returned as UTF-8?
I'm a bit clueless here.

Thanx in advance

Posted: Tue Aug 15, 2006 5:58 am

there are two settings to care of in this case:
$soapclient->soap_defencoding that should be set to 'UTF-8' or 'ISO-8859-1'
decode_utf8 property that can be true or false

Posted: Tue Aug 15, 2006 6:04 am

i assume the first suggestion depends on the client (which is not on my end). decode_utf8 is set to true here which is the standard setting. To check myself i ran the wsdl page in firefox which produced an ISO-8859-1 output.
so basically i can tell the client that the encoding depends on his setup?

Posted: Tue Aug 15, 2006 7:11 pm

ahh right,
I missed the point that it's server..

Server handles encoding like shown below:

Code:

      if(isset($this->headers['content-type']) && strpos($this->headers['content-type'],'=')){
         $enc = str_replace('"','',substr(strstr($this->headers["content-type"],'='),1));
         if(eregi('^(ISO-8859-1|US-ASCII|UTF-8)$',$enc)){
            $this->xml_encoding = strtoupper($enc);
         } else {
            $this->xml_encoding = 'US-ASCII';
         }
      } else {
         // should be US-ASCII for HTTP 1.0 or ISO-8859-1 for HTTP 1.1
         $this->xml_encoding = 'ISO-8859-1';
      }

So if client specifies either ISO-8859-1 or US-ASCII or UTF-8 by issuing Content-Type http header, it will be used, otherwise server falls back to ISO-8859-1.

Posted: Wed Aug 16, 2006 1:43 am

For testing i changed the fallback branch to UTF-8, even AFTER the else part so it would definately end up with UTF-8 no matter what:

Code:

      if(isset($this->headers['content-type']) && strpos($this->headers['content-type'],'=')){
         $enc = str_replace('"','',substr(strstr($this->headers["content-type"],'='),1));
         if(eregi('^(ISO-8859-1|US-ASCII|UTF-8)$',$enc)){
            $this->xml_encoding = strtoupper($enc);
         } else {
            $this->xml_encoding = 'US-ASCII';
         }
      } else {
         // should be US-ASCII for HTTP 1.0 or ISO-8859-1 for HTTP 1.1
         //$this->xml_encoding = 'ISO-8859-1';
         // the following did not seem to have any effect...
         $this->xml_encoding = 'UTF-8';
      }
      // so i put it here so it would end up with UTF-8
      $this->xml_encoding = 'UTF-8';

but nothing helped, the xml reply still shows

Code:

<?xml version="1.0" encoding="ISO-8859-1"?>

i even replaced all occurences of

Code:

$this->xml_encoding = 'ISO-8859-1';

with

Code:

$this->xml_encoding = 'UTF-8';

to provocate another behaviour but no success Sad

btw: appreciate ur support Smile

Posted: Wed Aug 16, 2006 2:01 am

I also set soap_defencoding to UTF-8 and called the WSDL with my own nusoap client:

Code:

/**
* charset encoding for outgoing messages
*
* @var string
* @access public
*/
//var $soap_defencoding = 'ISO-8859-1';
var $soap_defencoding = 'UTF-8';

result:

Code:

Request

POST /soap.php?wsdl HTTP/1.0
Host: soap.myserver
User-Agent: NuSOAP/0.7.2 (1.94)
Content-Type: text/xml; charset=UTF-8
SOAPAction: ""
Content-Length: 554

(...)

Response

HTTP/1.1 200 OK
Date: Thu, 17 Aug 2006 06:55:51 GMT
Server: Apache
X-Powered-By: PHP/5.1.4-0.0bpo1
Connection: close
Content-Type: text/xml; charset=ISO-8859-1

Posted: Wed Aug 16, 2006 12:15 pm

Maybe it will be faster, if you setup NuSphere PHPED with php debugger and step through your own and nusoap code to see how everything is handled?
I think it's not really hard. Please let me know if there are any problems with this "approach" Smile

Posted: Thu May 03, 2007 7:51 am

esto funciono para mi

hay que reemplazar todas las entradas de xml_encoding como figura a continuacion:

"xml_encoding =" (sin comillas) por "xml_encoding = 'UTF-8'; //" (nuevamente, sin comillas)