List directories containing Unicode characters on Windows

Question

I am using Windows 2003 to write some PHP code. I use XAMPP Portable (copy to D:). The problem:

$path = 'D:\ebooks';
$all_file = scandir($path);
foreach ($all_file as $file) {
  if (is_dir("$path/$file") && $file != '.' && $file != '..') {
    echo $file . "<br />\n";
  }
}

When I call the script (with browser), I didn't see any directories (within D:\ebooks) containing a Unicode character (I tested with Vietnamese, Japanese, Chinese, Czech).

But if I remove is_dir("$path/$file"), the directories display with many strange characters and many ??? characters.

How can I solve the problem?

What do you mean by "unicode name" ? you mean special characters like Ô or ü ? — Abhay Chaware, Apr 04 '12 at 09:28
The "many strange characters" part is possibly related to your encoding of the browser output. — cmbuckley, Apr 04 '12 at 16:36

Evert · Accepted Answer · 2012-04-04T13:55:28.667

6

Unfortunately, there are a lot of bugs related to PHP access to a windows filesystem. While Windows does store the filenames as UTF-16, PHP's internals use the much older ANSI api's. So it's best to only do stuff with filenames that are in the ascii range, or switch to a different operating system.

edited Apr 04 '12 at 13:55

answered Apr 04 '12 at 09:32

Evert

93,428
18
118
189

1

@cbuckley: Normally I'd discourage people from saying that, but in this case it's just fact that PHP on windows has really shitty filesystem support. – Evert Apr 04 '12 at 17:12
i have this problem in linux php 7.1. please check link: https://bugs.php.net/bug.php?id=74958 – user3770797 Jul 21 '17 at 17:36

score 0 · Answer 2 · edited May 23 '17 at 10:30

This might help with the Czech, it won't help with CJK (still looking for that myself). Buggy "Rationale": UTF-8 is ok, as long as it's mappable to the windows character set.

PHP is unpleasantly clear about this:

scandir() and readdir() do not support Unicode filenames

http://bugs.php.net/bug.php?id=34574

«PHP does not support unicode operations until PHP6.»

written in September 2005.

You could also try to grab the dos short name equivalent by using system('dir /X'), parse the results for short and long names and work from there. Display long names, actually accessing (for stat, timestamp and file open of course) with the short names...

List directories containing Unicode characters on Windows

2 Answers2

Linked