Re: serialize() and special ANSI characters

Roman Neuhauser wrote:

# ylacan@teicam.com / 2007-02-19 15:56:15 +0100:

I'm just curious to find out if I'm the only person to have bumped into this kind of issue with serialize/unserialize. When I try and serialize an array containing a string value with the "?" character (alt+241 ASCII) such as : "120GB 2X512MB 15.4IN DVD?RW VHP FR" The resulting serialized array is truncated. ie. I would obtain : "a:17:{i:0;s:1:"A";i:1;s:7:"TOSHIBA";i:2;s:4:"3740";i:3;s:7:"404D862";i:4;s:31:"SATELLITE A100-044 CD/T2060-1.6";i:5;s:35:"120GB 2X512MB 15.4IN DVD" As you can see serialization seems to stall as soon as the "?" character shows up.

I don't see the effect you're mentioning, but this is on FreeBSD, so perhaps there's a problem on Windows. I don't really believe it, though, and would hazard a guess you're running into display problems with your browser. Another thing to mention is that ASCII only goes up to 127. You may desire for ASCII 241 to mean +/-, but n-tilde is also a popular interpretation... Use a character set which actually includes +/- as a single character, and an encoding that can handle that charset.

I'm probably just running into so encoding issues with the CSV files I'm using on which I have no control whatsoever.

The following test fails in 5.2.1: <?php class serializeASCII241 extends Tence_TestCase {

    function testTruncates()
    {
        return $this->assertEquals(
            "120GB 2X512MB 15.4IN DVD",
            serialize("120GB 2X512MB 15.4IN DVD" . chr(241) . "RW VHP  FR")
        );
    }

} ?>

I'd just like to point out that your test will fail systematically unless you use something like this :

     function testTruncates()
     {

$string = "120GB 2X512MB 15.4IN DVD" . chr(241) . "RW VHP FR";

        return $this->assertEquals(
         "s:" . strlen($string) . ":\"120GB 2X512MB 15.4IN DVD",
         serialize("120GB 2X512MB 15.4IN DVD" . chr(241) . "RW VHP  FR")
         );
     }