serializerclass-general — general discussion

[serializerclass-general] the correct way (I think), test case

[Jamie P <ja...@e-...>]

Hi,

I think I found the correct way to calculate string length for serialising
and deserialising, both the code in the CVS and the code I previously sent
you don't work properly.

In the PHP manual it says Unicode is converted to UTF-8 as follows :


bytes bits representation
1 7 0bbbbbbb
2 11 110bbbbb 10bbbbbb
3 16 1110bbbb 10bbbbbb 10bbbbbb
4 21 11110bbb 10bbbbbb 10bbbbbb 10bbbbbb

It seems that Flash only supports Unicode characters up to 16 bits so we
will have to ignore 4 byte UTF-8. We should probably throw an error if
someone tries to send 4 byte but someone else can work on that one.

Here is the code that I have developed and tested with relevant test cases.

	private function getCStringLength ():Number {
		var colon:Number = this.buffer.indexOf (":", 3); // index of second colon
		var ss:Number = colon+2; // index of start of string
		var len:Number = parseInt (this.buffer.substr (2, colon - 2));
		var i:Number;
		var j:Number = len; //char length of string to be calculated
		var c:Number;
		var cstr = this.buffer;
		for (i = ss; i < (ss + j); i++)
		{
			c=cstr.charCodeAt(i);
			if (c<128) {
				//do nothing
			}else if (c<2048) {
				j = j-1;
			} else {
				j = j-2;
			}
		}
		return j;
	};	// getCStringLength



and used to serialize data :

	private function calcLength(struct:String)
	{
		var c;
		var result=0;
		var l = struct.length;
		for (var i=0; i < l; i++)
		{
			c = (struct.charCodeAt(i));
			if(c<128) {
				result += 1;
			} else if (c<2048) {
				result += 2;
			} else {
				result += 3;
			}
		}
		return result;
	}

This seems to be able to serialize and unserialise all unicode characters
supported by Flash properly and construct / reconstruct arrays and strings
which can be unserialised and serialised by php .

Check out the zip file for my test case here :

http://jamiep.org/mbcs/test.zip

and see it in action here :

http://jamiep.org/mbcs/serialtest.html

ALso you may be interested in the way I found to send data to the server :
override XML.toString(). It seems if you make XML.toString() return a
string and then use XML.sendAndLoad() or XML.send() that exactly the
string you return is sent. If you use the XML constructor then XML
converts & to &amp; but not if you override XML.toString(). Old news maybe?


Jamie

[serializerclass-general] mbcs fix; oh you fixed it already, which fix is right? - file

[Jamie P <ja...@e-...>]

Here is the fixed serialise class file in full, with some more minor  
changes.

[serializerclass-general] mbcs fix; oh you fixed it already, which fix is right?

[Jamie P <ja...@e-...>]

Hi all,

While reworking PHPobject I came across a bug in Alessandro's serializer  
class multibyte support.

I have tried to fix the deserialize function which wasn't handling MB  
strings properly. Here are the main changes (- marks the old code + the  
new):

+	private function getCStringLength ():Number {
+		var colon:Number = this.buffer.indexOf (":", 3); // index of second  
colon
+		var ss:Number = colon+2; // index of start of string
+		var len:Number = parseInt (this.buffer.substr (2, colon - 2));
  		var i:Number;
-		var j:Number = len;
+		var j:Number = len; //char length of string to be calculated
+		var c:Number;
  		var cstr = this.buffer;
-		for (i = 0; i < j; i++)
+		for (i = ss; i < (ss + j); i++)
  		{
-			if (cstr.charCodeAt (i + 5) > 128)
-			{
-				j = j - 2;
+			c=cstr.charCodeAt(i);
+
+			if (c<128) {
+				//do nothing
+			}else if (c<1024) {
+				j = j-1;
+			} else if (c<32768) {
+				j = j-2;
+			} else if (c<2097152) {
+				j = j-3;
  			}
  		}
  		return j;
-	};	// getCStringLenght
+	};	// getCStringLength

I am now hoping to get PHPobject working with UTF-8 without all the  
escaping and urlencoding that is being used now, which I think is largely  
or probably completely unnecessary.

I see that there is already a fix for this issue in the CVS! But my fix is  
different to yours. My fix does the reverse of your serialize function.  
Which is right I wonder?? I'm not sure how utf-8 is encoded in Flash.  
According to the Flash docs then char code can only go to 65535.

Jamie

[serializerclass-general] SeralizerClass and Special Characters: "æ ø å"

[søren j. <dem...@ho...>]

Hi all!

I've recently discovered the SerializerClass, and have begun experimenting 
with it ;)

Though all seems to work fine i'm having trouble unserializing data from 
flash in php,
that contains special characters like: "æ ø å"

My PHP looks the following:

<?
error_reporting(E_ALL);
/*
echo "<br>showing post<br>";
var_dump($_POST);
echo "<br>showing get<br>";
var_dump($_GET);
$a = $_POST['xmldata'];

echo "<br>showing name=\"xmldata\"<br>";
$b = $a['xmldata'];
echo $a['xmldata'];
$c = $b['Shippinginfo'];
var_dump($b['Shippinginfo']);
$d = $c['Name'];

echo get_object_vars($c);


$xml = unserialize(urldecode(stripslashes($_REQUEST['xmldata'])));

var_dump($xml);

print_r($_REQUEST['xmldata']);
*/

class xmldata {
}

$xml = unserialize(urldecode(stripslashes($_REQUEST['xmldata'])));

var_dump($xml);
?>

Any suggestions to what i'm doing wrong?

Secondly is this the way to go, if i want to transport large amounts of data 
in the area of 300-400 Kb's of data? Or is it better to use the AMF format?

Regards

Søren Jepsen

_________________________________________________________________
Få alle de nye og sjove ikoner med MSN Messenger http://messenger.msn.dk