Description:
------------
PHP is assuming character encoding from external executed svn client, as ASCII.

Even when external program returns ISO-8859-1 encoded string, PHP "parses" the encoded string as ASCII, expanding accented characters as literal string form and not their binary form.

For example: 
an output like "Acentua??o" turns to be a string in literal form "Acentua?\195?\167?\195?\163o/".

Reproduce code:
---------------
Import some accented file or folders into a subversion repository. Is it possible to convert the output to utf-8 using the command bellow:

# svn list 'file:////home/svn/herbert/' | iconv -tutf-8

But not when from PHP:

<?php
$cmd = "svn list 'file:////home/svn/herbert/'";
$out = shell_exec($cmd);
$res = unpack('c*', $out);
var_dump($res);
?>

var_dump reports:

array(29) {
  [1]=>
  int(65)
  [2]=>
  int(99)
  [3]=>
  int(101)
  [4]=>
  int(110)
  [5]=>
  int(116)
  [6]=>
  int(117)
  [7]=>
  int(97)
  [8]=>
  int(63)
  [9]=>
  int(92)
  [10]=>
  int(49)
  [11]=>
  int(57)
  [12]=>
  int(53)
  [13]=>
  int(63)
  [14]=>
  int(92)
  [15]=>
  int(49)
  [16]=>
  int(54)
  [17]=>
  int(55)
  [18]=>
  int(63)
  [19]=>
  int(92)
  [20]=>
  int(49)
  [21]=>
  int(57)
  [22]=>
  int(53)
  [23]=>
  int(63)
  [24]=>
  int(92)
  [25]=>
  int(49)
  [26]=>
  int(54)
  [27]=>
  int(51)
  [28]=>
  int(111)
  [29]=>
  int(47)
}

So it's not possible to convert the string to other character set, since it's invalid.

Expected result:
----------------
It's expected to PHP store the string as it's original binary format.

array(10) {
  [1]=>
  int(65)
  [2]=>
  int(99)
  [3]=>
  int(101)
  [4]=>
  int(110)
  [5]=>
  int(116)
  [6]=>
  int(117)
  [7]=>
  int(97)
  [8]=>
  int(-25)
  [9]=>
  int(-29)
  [10]=>
  int(111)
}