explode problem with delimeter chr 178

[tekken00]

I am trying to split a string by a non printable ascii char 178 and it does kind of work
but explode() will keep delimeter in splitted strings if delimiter is non printable, meaning that
every array element will contain chr 178. I hope it does not sound too crazy because
when I did trace this behaivor I could not believe it myself.

$Delim=chr(178);
$Data=explode($Delim,Content);

CentOS/PHP 5.2.10 (and same problem on Debian)

[bradgrafelman]

Doesn't seem to be the case for me:

$str = '123' . chr(178) . '1234' . chr(178) . '1';

var_dump( explode(chr(178), $str) );

produced:

array(3) {
  [0]=>
  string(3) "123"
  [1]=>
  string(4) "1234"
  [2]=>
  string(1) "1"
}

The string lengths would seem to indicate that the chr(178) character isn't in the string.

Also, chr(178) is not a non-printable character - it's quite printable (it's a shaded rectangular box).

[johanafm]

Either you're constant "Content" is not containing the data you believe it does, or perhaps you get issues due to character encoding issues.

ASCII contains 128 characters, i.e. chr(0) to char(127). chr(178) is not ascii, since it's part of an ascii extension, and as such depends on the encoding used. However, whatever character chr(178) is, it will be removed if used as delimiter.

[tekken00]

after more extensive digging I found out the string is coming from .net remote app (not website) and encoding is most likely the problem since .net uses utf16. I got the source and verified that chr 178 is properly used as delimeter however .net app can not be changed so it has to be somehow fixed on php side. Also the grabage chr left in splitted strings is actually 198 not 178
which makes more like to be an encoding issue (or php decoding actually in this case)

[johanafm]

Well, the UTF-16 character is most likely the byte sequence chr(198) chr(178), which means you should explode on this sequence rather than just chr(178). If you need output in anything other than utf-16, maybe iconv can help you